Anda di halaman 1dari 101

Expansive Measures

Publicaes Matemticas

Expansive Measures

Carlos A. Morales
UFRJ
Vctor F. Sirvent
Universidad Simon Bolivar

29o Colquio Brasileiro de Matemtica

Copyright 2013 by Carlos A. Morales e Vctor F. Sirvent

Impresso no Brasil / Printed in Brazil


Capa: Noni Geiger / Srgio R. Vaz

29o Colquio Brasileiro de Matemtica

Anlise em Fractais Milton Jara


Asymptotic Models for Surface and Internal Waves - Jean-Claude Saut
Bilhares: Aspectos Fsicos e Matemticos - Alberto Saa e Renato de S
Teles
Controle timo: Uma Introduo na Forma de Problemas e Solues Alex L. de Castro
Eigenvalues on Riemannian Manifolds - Changyu Xia
Equaes Algbricas e a Teoria de Galois - Rodrigo Gondim, Maria
Eulalia de Moraes Melo e Francesco Russo
Ergodic Optimization, Zero Temperature Limits and the Max-Plus
Algebra - Alexandre Baraviera, Renaud Leplaideur e Artur Lopes
Expansive Measures - Carlos A. Morales e Vctor F. Sirvent
Funes de Operador e o Estudo do Espectro - Augusto Armando de
Castro Jnior
Introduo Geometria Finsler - Umberto L. Hryniewicz e Pedro A. S.
Salomo
Introduo aos Mtodos de Crivos em Teoria dos Nmeros - Jlio
Andrade
Otimizao de Mdias sobre Grafos Orientados - Eduardo Garibaldi e
Joo Tiago Assuno Gomes

ISBN: 978-85-244-0360-6

Distribuio: IMPA
Estrada Dona Castorina, 110
22460-320 Rio de Janeiro, RJ
E-mail: ddic@impa.br
http://www.impa.br

Contents
Preface

iii

1 Expansive measures
1.1 Denition and examples . . . . . . . . . .
1.2 Expansive invariant measures . . . . . . .
1.3 Equivalences . . . . . . . . . . . . . . . .
1.4 Properties . . . . . . . . . . . . . . . . . .
1.5 Probabilistic proofs in expansive systems .
1.6 Exercices . . . . . . . . . . . . . . . . . .

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

1
1
6
11
15
22
25

2 Finite expansivity
2.1 Introduction . . . . .
2.2 Preliminaries . . . .
2.3 n-expansive systems
2.4 The results . . . . .
2.5 Exercices . . . . . .

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

27
27
27
32
37
38

.
.
.
.
.
.

40
40
41
46
49
56
57

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

3 Positively expansive measures


3.1 Introduction . . . . . . . . . .
3.2 Denition . . . . . . . . . . .
3.3 Properties . . . . . . . . . . .
3.4 Applications . . . . . . . . . .
3.5 The smooth case . . . . . . .
3.6 Exercices . . . . . . . . . . .
i

.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

ii
4 Measure-sensitive maps
4.1 Introduction . . . . . . .
4.2 Measure-sensitive spaces
4.3 Measure-sensitive maps
4.4 Aperiodicity . . . . . . .
4.5 Exercices . . . . . . . .
Bibliography

CONTENTS

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

60
60
61
64
69
76
79

Preface
It is customary to say that a given phenomenum is chaotic if it cannot
be predicted. This is what currently occurs in many circuntances like
in weather prediction, particle behavior in physic or nancial marked.
But what the meanning of predictiblity is? A simple manner to answer this question is to model the given phenomenum as the trajectories of a dynamical system and, then, reinterpret the predictibility as
the knowledgement of where trajectories go. For instance, in weather
prediction or particle behavior or nancial marked it is known that
nearby initial conditions can produce very dierent outputs thus characterizing a very high degree of unpredictibility. Such a situation is
easily described in dynamics with the notion of sensitivity to initial
conditions in which every face point can be approached by points for
which the corresponding trayectories eventually separate in the future
(or in the past for invertible systems). The worst scenario appears
precisely when the trajectory of every nearby point separate from
the initial one, and this is what is commonly denominated as expansive system. In these terms expansivity manifests the most chaotic
scenario in which predictions may have no sense at all.
The rst researcher who considered the expansivity in dynamics
was Utz in his seminal paper [86]. Indeed, he dened the notion
of unstable homeomorphisms (nowadays known as expansive homeomorphisms [39]) and studied their basic properties. Since then an extensive literature about these homeomorphisms has been developed.
For instance, [90] proved that the set of points doubly asymptotic
to a given point for expansive homeomorphisms is at most countable.
Moreover, a homeomorphism of a compact metric space is expansive
if it does in the complement of nitely many orbits [91]. In 1972 Sears
iii

iv

PREFACE

proved the denseness of expansive homeomorphisms with respect to


the uniform topology in the space of homeomorphisms of a Cantor
set [80]. An study of expansive homeomorphisms using generators
is given in [20]. Goodman [38] proved that every expansive homeomorphism of a compact metric space has a (nonnecessarily unique)
measure of maximal entropy whereas Bowen [11] added specication
to obtain unique equilibrium states. In another direction, [76] studied
expansive homeomorphisms with canonical coordinates and showed
in the locally connected case that sinks or sources cannot exist. Two
years later, Fathi characterized expansive homeomorphisms on compact metric spaces as those exhibiting adapted hyperbolic metrics
[34] (see also [78] or [30] for more about adapted metrics). Using this
he was able to obtain an upper bound of the Hausdor dimension and
upper capacity of the underlying space using the topological entropy.
In [54] it is computed the large deviations of irregular periodic orbits
for expansive homeomorphisms with the specication property. The
C 0 perturbations of expansive homeomorphisms on compact metric
spaces were considered in [24]. Besides, the multifractal analysis of
expansive homeomorphisms with the specication property was carried out in [84]. We can also mention [23] in which it is studied a new
measure-theoretic pressure for expansive homeomorphisms.
From the topological viewpoint we can mention [67] and [74] proving the existence of expansive homeomorphisms in the genus two
closed surface, the n-torus and the open disk. Analogously for compact surfaces obtained by making holes on closed surfaces dierent
from the sphere, projective plane and Klein bottle [51]. In [46] it was
proved that there are no expansive homeomorphisms of the compact
interval, the circle and the compact 2-disk. The same negative result
was obtained independently by Hiraide and Lewowicz in the 2-sphere
[42], [59]. Ma
ne proved in [62] that a compact metric space exhibiting expansive homeomorphisms must be nite dimensional and, further, every minimal set of such homeomorphisms is zero dimensional.
Previously he proved that the C 1 interior of the set of expansive diffeomorphisms of a closed manifold is composed by pseudo-Anosov
(and hence Axiom A) dieomorphisms. In 1993 Vieitez [87] obtained
results about expansive homeomorphisms on closed 3-manifolds. In
particular, he proved that the denseness of the topologically hyperbolic periodic points does imply constant dimension of the stable and

v
unstable sets. As a consequence a local product property is obtained
for such homeomorphisms. He also obtained that expansive homeomorphisms on closed 3-manifolds with dense topologically hyperbolic
periodic points are both supported on the 3-torus and topologically
conjugated to linear Anosov isomorphisms [88].
In light of these results it was natural to consider another notions
of expansiveness. For example, G-expansiveness, continuouswise and
pointwise expansiveness were dened in [29], [50] and [75] respectivelly. We also have the entropy-expansiveness introduced by Bowen
[10] to compute the metric and topological entropies in a large class
of homeomorphisms.
In this monograph we will consider a notion of expansiveness,
located in between sensitivity and expansivity, in which Borel probability measures will play fundamental role. Indeed, we say that is
an expansive measure of a homeomorphism f if the probability of two
orbits remain close each other up to a prexed radius is zero. Analogously, for continuous maps, we dene positively expansive measure
by considering positive orbits instead. The corresponding concepts
for certain topological spaces (e.g. uniform spaces) likewise ows or
topological group actions have been considered elsewhere [22], [66].
These concepts are closely related (and sometimes equivalent to)
the concepts of pairwise sensitivity [27] and the -sensitivity [44] in
which the sensitivity properties of these systems are emphasized.
Here we give emphasize not in the sensitivity but, rather, in the
expansivity properties of these systems.
In Chapter 1 we will give the precise denition of expansive measures for homeomorphisms f as well as some basic properties closely
related to the expansive systems. For instance, we characterize the
expansive measures as those for which the diagonal is almost invariant for f f with respect to the product measure 2 . In addition, we
prove that the set of heteroclinic points has measure zero with respect
to any expansive measure. In particular, the set of periodic orbits for
these homeomorphisms is also of measure zero for such measures. We
also prove that there are expansive measures for homeomorphisms in
any compact interval and, in the circle, we prove that they exists
solely for the the Denjoy maps. As an application we obtain probabilistic proofs of some result of expansive systems.
In Chapter 2 we will analyze the n-expansive systems which rep-

vi

PREFACE

resent a particular (an interesting) example of nonexpansive systems


for which every non-atomic Borel measure is expansive.
In Chapter 3 we study the class of positively expansive measures
and prove that every ergodic invariant measure with positive entropy
of a continuous map on a compact metric space is positively expansive. We use this property to prove, for instance, that the stable
classes have measure zero with respect to any ergodic invariant measure with positive entropy. Moreover, continuous maps which either
have countably many stable classes or are Lyapunov stable on their
recurrent sets have zero topological entropy. We also apply our results
to the Li-Yorke chaos.
Finally, in Chapter 4, we will extend the notion of expansivity to
include measurable maps on measure spaces. Indeed, we study countable partitions for measurable maps on measure spaces such that for
all point x the set of points with the same itinerary of x is negligible. We prove that in non-atomic probability spaces every strong
generator [69] satises this property but not conversely. In addition,
measurable maps carrying partitions with this property are aperiodic and their corresponding spaces are non-atomic. From this we
obtain a characterization of nonsingular countable to one mappings
with these partitions on non-atomic Lebesgue probability spaces as
those having strong generators. Furthermore, maps carrying these
partitions include the ergodic measure-preserving ones with positive
entropy on probability spaces (thus extending a result by Cadre and
Jacob [27]). Applications of these results will be given. At the end
of each chapter we include some exercices whose diculty was not
estimated. Some basics of dynamical systems, ergodic and measure
theory will be recommendable for the comprension of this text.

September 2012
UFRJ, USB

C. A. M. & V. F. S.
Rio de Janeiro, Caracas.

Acknowledgments
The authors want to thank the Instituto de Matematica Pura e
Aplicada (IMPA) and the Sim
on Bolvar University for their kindly
hospitality. They also thank their colleagues professors Alexander Arbieto, Dante Carrasco-Olivera, Jose Carlos Martin-Rivas and Laura
Senos by the invaluable mathematical conversations.
C.A.M. was partially supported by FAPERJ, CAPES, CNPq,
PRONEX-DYN. SYS. from Brazil and the Sim
on Bolvar University
from Venezuela.

vii

Chapter 1

Expansive measures
1.1

Denition and examples

In this section we introduce the denition of expansive measures for


homeomorphisms and present some examples. To motivate let us
recall the concept of expansive homeomorphism.
Denition 1.1. A homeomorphism f : X X of a metric space
X is expansive if there is > 0 such that for every pair of dierent
points x, y X there is n Z such that d(f n (x), f n (y)) > .
An important remark is given below.
Remark 1.2. Equivalently, f is expansive if there is > 0 such that
(x) = {x} for all x X where
(x) = {y X : d(f i (x), f i (y)) , i Z}.
(Notation f (x) will indicate dependence on f .)
This denition suggests further notions of expansiveness involving
a given property (P) of the closed sets in X. More precisely, we say
that f is expansive with respect to (P) if there is > 0 such that
(x) satises (P) for all x X.
For example, a homeomorphism is expansive it is expansive with
respect to the property of being a single point. Analogously, it is
1

[CAP. 1: EXPANSIVE MEASURES

h-expansive (c.f. [10]) if it is expansive with respect to the property


of being a zero entropy set. In this vein it is natural to consider the
property of having zero measure with respect to a given Borel probability measure of X. By Borel measure we mean a non-negative
-additive function dened in the Borel -algebra of X which is
non-zero in the sense that (X) > 0.
Denition 1.3. A expansive measure of homeomorphism f : X X
of a metric space X is a Borel measure for which there is > 0
such that ( (x)) = 0 for all x X. The constant will be referred
to as an expansivity constant of .
Let us present some examples related to this denition. Recall
that a Borel measure of a metric space X is a probability if (X) = 1
and non-atomic if ({x}) = 0 for all x X.
Example 1.4. Every expansive measure is non-atomic. Therefore,
every metric space carrying homeomorphisms with expansive (probability) measures also carries a non-atomic Borel (probability) measure.
In the converse direction we have the following relation between
expansive homeomorphisms and expansive measures for homeomorphisms.
Example 1.5. If f : X X is an expansive homeomorphism of
a metric space X, then every non-atomic Borel measure of X (if it
exists) is an expansive measure of f . Moreover, all such measures
have a common expansivity constant.
Example 1.5 motivates the question whether a homeomorphism
is expansive if it satises that every non-atomic Borel measure (if it
exists) is expansive with a common expansivity constant. We shall
give a partial positive answer based on the following denition (closely
related to that of expansive homeomorphism).
Denition 1.6. A homeomorphism f : X X of a metric spaces X
is countably-expansive if there is > 0 such that (x) is countable,
x X.

[SEC. 1.1: DEFINITION AND EXAMPLES

Clearly, every expansive homeomorphism is countably-expansive


but not conversely (as we shall see in Chapter 2). In addition, every countably-expansive homeomorphism satises that all non-atomic
Borel probability measures (if they exist) are expansive with common
expansivity constant. The following result proves the converse of this
last assertion for Polish metric spaces, i.e., metric spaces which are
both complete and separable.
Proposition 1.7. The following properties are equivalent for every
homeomorphism f : X X of a Polish metric space X:
1. f is countably-expansive.
2. All non-atomic Borel probability measures of X (if they exit)
are expansive with a common expansivity constant.
Proof. By the previous discussion we only have to prove that (2) implies (1). Suppose by contradiction that all non-atomic Borel probability measures are expansive measures with a common expansivity
constant (say ) but f is not countably-expansive. Then, there is
x X such that (x) is uncountable. Since (x) is also a closed
subset of X which is a Polish metric space, we have that (x) is
a Polish metric space too. Then, we can apply a result in [73] (e.g.
Theorem 8.1 p. 53 in [72]) to obtain a non-atomic Borel probability
of X supported on (x). For such a measure we would obtain
( (x)) = 1 a contradiction.
In light of this proposition it is natural to ask what can happen
if we still assume that all non-atomic Borel probability measure (if it
exist) are expansive but without assuming that they have a common
expansivity constant.
This question emphasizes the role of metric spaces for which there
are non-atomic Borel probability measures. For the sake of convenience we call these spaces non-atomic metric spaces. The aforementioned result in [73] (stated in Theorem 8.1 p. 53 in [72]) implies that
every uncountable Polish metric space is a non-atomic metric space.
This includes the compact metric space containing perfect subsets
[55]. Every non-atomic metric space is uncountable.
Another related denition is as follows.

[CAP. 1: EXPANSIVE MEASURES

Denition 1.8. A homeomorphism f : X X of a non-atomic


metric space X is measure-expansive if every non-atomic Borel probability measure is expansive for f .
It is clear that every expansive homeomorphism of a non-atomic
metric space is measure-expansive. Moreover, as discussed in Example 1.5, every countably-expansive homeomorphism of a non-atomic
metric space is measure-expansive. Although we obtain in Example
3.44 that there are measure-expansive homeomorphisms of compact
non-atomic metric spaces which are not expansive, we dont know
any example of one which is not countably-expansive (see Problem
1.46). Some dynamical consequences of measure-expansivity resembling expansivity will be given later on.
Further examples of homeomorphisms without expansive measures can be obtained as follows. Recall that an isometry of a metric
space X is a map f : X X satisfying d(f (x), f (y)) = d(x, y) for
all x, y X.
Example 1.9. Every isometry of a separable metric space has no expansive measures. In particular, the identity map in these spaces (or
the rotations in R2 or translations in Rn ) are not measure-expansive
homeomorphisms.
Proof. Suppose by contradiction that there is a an expansive measure
for some isometry f of a separable metric space X. Since f is an
isometry we have (x) = B[x, ], where B[x, ] denotes the closed
-ball around x. If is an expansivity constant, then (B[x, ]) =
( (x)) = 0 for all x X. Nevertheless, since X is separable (and
so Lindel
of), we can select a countable covering {C1 , C2 , , Cn , }
of X by closed subsets such that
for all n thereis xn X such that


Cn B[xn , ]. Thus, (X) n=1 (Cn ) n=1 (B[xn , ]) = 0


which is a contradiction. This proves the result.
Example 1.10. Endow Rn with a metric space with the Euclidean
metric and denote by Leb the Lebesgue measure in Rn . Then, Leb is
an expansive measure of a linear isomorphism f : Rn Rn if and
only if f has eigenvalues of modulus less than or bigger than 1.
Proof. Since f is linear we have (x) = (0) + x thus Leb( (x)) =
Leb( (0)) for all x Rn and > 0. If f has eigenvalues of modulus

[SEC. 1.1: DEFINITION AND EXAMPLES

less than or bigger than 1, then (0) is contained in a proper subspace of Rn which implies Leb( (0)) = 0 thus Leb is expansive.
Example 1.11. As we shall see later, a homeomorphism of a compact interval has no expansive measures. In the circle the sole homeomorphisms having such measures are the Denjoy ones.
Recall that a subset Y X is invariant if f 1 (Y ) = Y .
Example 1.12. A homeomorphism f has an expansive measure if
and only if there is an invariant borelian set Y of f such that the
restriction f /Y has an expansive measure.
Proof. We only have to prove the only if part. Assume that f /Y
has an expansive measure . Fix > 0. Since Y is invariant we
f /Y
have either f/2 (x) Y = or f/2 (x) Y (y) for some
f /Y

y Y . Therefore, either f/2 (x) Y = or (f/2 (x)) ( (y))


for some y Y where is the Borel probability of X dened by
(A) = (A Y ). From this we obtain that for all x X there is
f /Y
y Y such that (f/2 (x)) ( (y)). Taking as an expansivity
constant of f /Y we obtain (f/2 (x)) = 0 for all x X thus is
expansive with expansivity constant /2.
The next example implies that the property of having expansive
measures is a conjugacy invariant. Given a Borel measure in X
and a homeomorphism : X Y we denote by () the pullback
of dened by ()(A) = (1 (A)) for all borelian A.
Example 1.13. Let be an expansive measure of a homeomorphism
f : X X of a compact metric space X. If : X Y is a
homeomorphism of compact metric spaces, then () is an expansive
measure of f 1 .
Proof. Clearly is uniformly continuous so for all > 0 there is  > 0
such that f (y) (f (1 (y))) for all y Y . This implies

()(f
(y)) (f (1 (y))).


Taking as the expansivity constant of we obtain that  is an


expansivity constant of ().

[CAP. 1: EXPANSIVE MEASURES

For the next example recall that a periodic point of a homeomorphism (or map) f : X X is a point x X such that f n (x) = x for
some n N+ . The nonwandering set of f is the set (f ) of points
x X such that for every neighborhood U of x there is n N+ satisfying f n (U )U
= . Clearly a periodic point belongs to (f ) but not
every point in (f ) is periodic. If X = M is a closed (i.e. compact
connected boundaryless Riemannian) manifold and f is a dieomorphism we say that an invariant set H is hyperbolic if there are a cons
u
EH
tinuous invariant tangent bundle decomposition TH M = EH
and positive constants K, > 1 such that
Df n (x)/Exs Kn

and

m(Df n (x)/Exu ) K 1 n ,

for all x H and n IN (m denotes the co-norm operation in M ).


We say that f is Axiom A if (f ) is hyperbolic and the closure of
the set of periodic points.
Example 1.14. Every Axiom A dieomorphism with innite nonwandering set of a closed manifold has expansive measures.
Proof. Consider an Axiom A dieomorphism f of a closed manifold. It is well known that there is a spectral decomposition (f ) =
H1 Hk consisting of nitely many disjoint homoclinic classes
H1 , , Hk of f (see [40] for the corresponding denitions). Since
(f ) is innite we have that H = Hi is innite for some 1 i k.
As is well known f /H is expansive. On the other hand, H is compact
without isolated points since it is a homoclinic class. It follows from
Example 1.5 that f /H has an expansive measure, so, f also has by
Example 1.12.
We shall prove in the next section that every homeomorphism
with expansive measures of a compact metric space has uncountable
nonwandering set.

1.2

Expansive invariant measures

Let f : X X be a continuous map of a metric space X. We


say that a Borel measure of X is invariant if f = . In this

[SEC. 1.2: EXPANSIVE INVARIANT MEASURES

section we investigate the existence of expansive invariant measures


for homeomorphisms on compact metric spaces.
Indeed, every homeomorphism of a compact metric space carries
invariant measures, but not necessarily expansive measures (e.g. the
circle rotations). On the other hand, the homeomorphism f (x) = 2x
on the real line exhibits expansive probability measures (e.g. the
Lebesgue measure supported on the unit interval) but not expansive invariant probability measures. The result of this section will
show that the situation described in this example does not occur on
compact metric spaces. More precisely, we will show that every homeomorphism exhibiting expansive probability measures of a compact
metric space also exhibit expansive invariant probability measures.
We start with the following observation where f is assumed to be
a bijective map, namely,
(x, ) X R+ .

f ( (x)) = (f (x)),

Using it we obtain the elementary lemma below.


Lemma 1.15. Let f : X X be a homeomorphism of a metric
space X. If is an expansive measure with expansivity constant of
f , then so does f .
Proof. Applying the previous observation to f 1 we obtain
f ( (x)) = (f 1 ( (x)) = ( (f 1 (x))) = 0
for all x X.
Another useful observation is as follows. Given a bijective map
f : X X, x X, > 0 and n N+ we dene
V [x, n, ] = {y X : d(f i (x), f i (y)) , for all n i n},
i.e.,
V [x, n, ] =

n


f i (B[f i (x), ]).

i=n

(when necessary we write Vf [x, n, ] to indicate dependence on f .) It


is then clear that

V [x, n, ]
(x) =
nN+

[CAP. 1: EXPANSIVE MEASURES

and that V [x, n, ] V [x, m, ] for n m. Consequently,


( (x)) = lim (V [x, kl , ])
l

(1.1)

for every x X, > 0, every Borel probability measure of X, and


every sequence kl . From this we have the following lemma.
Lemma 1.16. Let f : X X be a homeomorphism of a metric
space X. A Borel probability measure is an expansive measure of
f if and only if there is > 0 such that
lim inf (V [x, n, ]) = 0,
n

for all x X.

We shall use this information in the following lemma.


Lemma 1.17. If f : X X is a homeomorphism of a metric space
X, then every invariant measure of f which is the limit (with respect
to the weak-* topology) of a sequence of expansive probability measures
with a common expansivity constant of f is expansive for f .
Proof. Denote by A = Cl(A)\Int(A) the closure of a subset A X.
Let be an invariant probability measure of f . As in the proof of
Lemma 8.5 p. 187 in [40] for all x X we can nd 2 < x < such
that
((B[x, x ])) = 0.
This allows us to dene
W [x, n] =

n


f i (B[f i (x), f i (x) ]),

(x, n) X N.

i=n

Since

< x < we can easily verify that





V x, n,
W [x, n] V [x, n, ],
(x, n) X N.
2

Moreover, as f (and so f i ) are homeomorphisms one has



 n

i
i
f (B[f (x), f i (x) ])
(W [x, n]) =
i=n

(1.2)

[SEC. 1.2: EXPANSIVE INVARIANT MEASURES


n


n


f i (B[f i (x), f i (x) ]) =


f i (B[f i (x), f i (x) ]) ,

i=n

i=n

and, since is invariant,


((W [x, n]))

(f i (B[f i (x), f i (x) ]) ) =

i=n
n

((B[f i (x), f i (x) ])) = 0,

i=n

proving
((W [x, n])) = 0,

(x, n) X N.

(1.3)

Now, suppose that is the weak-* limit of a sequence of expansive


probability measures n withoutcommon expansivity constant of f .
Clearly, is also a probability measure. Fix x X. Since each n is
a probability we have 0 m (W [x, n]) 1 for all n, m N. Then,
we can apply the Bolzano-Weierstrass Theorem to nd sequences
kl , rs for which the double limit
lim rs (W [x, kl ])

l,s

exists.
On the one hand, for xed l, using (1.3), n and well-known
properties of the weak-* topology (e.g. Theorem 6.1-(e) p. 40 in [72])
one has that the limit
lim rs (W [x, kl ]) = (W [x, kl ])

exists.
On the other hand, the second inequality in (1.2) and (1.1) imply
for xed s that
lim rs (W [x, kl ]) lim rs (V [x, kl , ]) = rs ( (x)) = 0.

Consequently, the limit


lim rs (W [x, kl ]) = 0

10

[CAP. 1: EXPANSIVE MEASURES

also exists for xed s.


From these assertions and well-known properties of double sequences one obtains
lim lim rs (W [x, kl ]) = lim lim rs (W [x, kl ]) = 0.

l s

s l

But (1.2) implies







lim inf V x, n,
lim V x, kl ,
lim (W [x, kl ])
n
l
l
2
2
and n together with (1.3) yields
lim (W [x, kl ]) = lim lim rs (W [x, kl ])
l s

so




lim inf V x, n,
=0
n
2

and then is expansive by Lemma 1.16.


Using these lemmas we obtain the following result.
Theorem 1.18. A homeomorphisms of a compact metric space has
an expansive probability measure if and only if it has an expansive
invariant probability measure.
Proof. Let be an expansive measure (with expansivity constant )
of a homeomorphism f : X X of a compact metric space X.
By Lemma 1.15 we have that fi is also an expansive probability
measure with expansivity constant (i Z). Therefore,
n =

n1
1 i
f ,
n i=0

n N+

is a sequence of expansive probability measures of f with common


expansivity constant . As X is compact there is a subsequence
nk such that nk converges to a Borel probability measure .
Since is clearly invariant we can apply Lemma 1.17 to this sequence
to obtain that is expansive.

11

[SEC. 1.3: EQUIVALENCES

A direct consequence of Theorem 1.18 is as follows. First of all


denote by supp() the support of a Borel measure . Given a metric
space X and a map f : X X we dene the omega-limit set of
x X,


(x) = y X : y = lim f nk (x) for some sequence nk .
k

The recurrent set of f is given by


R(f ) = {x X : x (x)}.
With these denitions we have the following corollary. Denote by
supp() the support of a Borel measure of a metric space X.
Corollary 1.19. The recurrent (and hence the nonwandering) sets
of every homeomorphism with expansive probability measures of a
compact metric space is uncountable.
Proof. Let f : X X be a homeomorphism with an expansive
probability measure of a compact metric space X. By Theorem
1.18 we can assume that is invariant, and so, supp() R(f ) by the
Poincare Recurrent Theorem. If R(f ) were countable we would have
(supp()) (R(f )) = 0 which is absurd thus R(f ) is uncountable.

1.3

Equivalences

In this section we present some equivalences for the expansivity of a


given measure. Hereafter all metric spaces X under consideration will
be compact unless otherwise stated. We also x a Borel probability
measure of X and a homeomorphism f : X X.
To start we observe an apparently weak denition of expansive
measure saying that is an expansive measure of f if there is > 0
such that ( (x)) = 0 for -a.e. x X. However, this denition
and the previous one are in fact equivalent by the following lemma.
Lemma 1.20. Let f : X X be a homeomorphism of a compact metric space X. Then, a Borel probability measure of X is
an expansive measure of f if and only if there is > 0 such that
( (x)) = 0 for -a.e. x X.

12

[CAP. 1: EXPANSIVE MEASURES

Proof. We only need to prove the if part. Let > 0 be such that
( (x)) = 0 for -a.e. x X. We shall prove that /2 is an
expansiveness constant of . Suppose by contradiction that it is
not so. Then, there is x0 X such that (/2 (x0 )) > 0. Denote
A = {x X : ( (x)) = 0} so (A) = 1. Since is a probability
measure we obtain A /2 (x0 )
= so there is y0 /2 (x0 ) such
that ( (y0 )) = 0.
Now, since y0 /2 (x0 ) we have /2 (x0 ) (y0 ). Indeed
d(f i (x), f i (x0 )) /2 (i N) implies
d(f i (x), f i (y0 )) d(f i (x), f i (x0 )) + d(f i (x0 ), f i (y0 ))
/2 + /2 = ,

i N

proving the assertion. It follows that (/2 (x0 )) ( (y0 )) = 0


which is a contradiction. This proves the result.
In particular, we have the following corollary.
Corollary 1.21. Let f : X X be a homeomorphism of a compact
metric space X. Then, a Borel probability measure is an expansive
measure of f if and only if there is > 0 such that ( (x)) = 0 for
all x supp().
A direct application of Lemma 1.16 is the following version of a
well-known property of the expansive homeomorphisms (see Corollary 5.22.1-(ii) of [89]).
Proposition 1.22. Let f : X X a homeomorphism and be a
Borel probability measure of a compact metric space X. If n Z\{0},
then is an expansive measure of f if and only if it is an expansive
measure of f n .
Proof. We can assume that n > 0. First notice that Vf [x, n m, ]
Vf n [x, m, ]. If is an expansive measure of f n is expansive then by
Lemma 1.16 there is > 0 such that for every x X there is a sequence mj such that (Vf n [x, mj , ]) 0 as j . Therefore
(Vf [x, n mj , ]) 0 as j yielding lim inf n (Vf [x, n, ]) =
0. Since x is arbitrary we conclude that is expansive with constant
.

[SEC. 1.3: EQUIVALENCES

13

Conversely, suppose that is an expansive measure of f with


constant . Since X is compact and n is xed we can choose 0 <  <
such that if d(x, y) , then d(f i (x), f i (y)) < for all n i n.
n
With this property one has f (x) f (x) for all x X thus is
an expansive measure of f n with constant .
One more equivalence is motivated by a well known condition for
expansiveness. Given metric spaces X and Y we always consider the
product metric in X Y dened by
d((x1 , y1 ), (x2 , y2 )) = d(x1 , x2 ) + d(y1 , y2 ).
If and are measures in X and Y respectively we denote by
their product measure in X Y . If f : X X and g : Y Y we
dene their product f g : X Y X Y ,
(f g)(x, y) = (f (x), g(y)).
Notice that f g is a homeomorphism if f and g are. Denote by
= {(x, x) : x X} the diagonal of X X.
Given a map g of a metric space Y we call an invariant set I
isolated if there is a compact neighborhood U of it such that
I = {z U : g n (z) U, n Z}.
As is well known, a homeomorphism f of X is expansive if and only
if the diagonal is an isolated set of f f (e.g. [4]). To express the
corresponding version for expansive measures we introduce the following denition. Let be a Borel probability measure of Y . We call
an invariant set I of g -isolated if there is a compact neighborhood
U of I such that
({z Y : g n (z) U, n Z}) = 0.
With this denition we have the following result in which we write
2 = .
Theorem 1.23. Let f : X X be a homeomorphism of a compact
metric space X. Then, a Borel probability measure of X is an
expansive measure of f if and only if the diagonal is a 2 -isolated
set of f f .

14

[CAP. 1: EXPANSIVE MEASURES

Proof. Fix > 0 and a -neighborhood U = {z X X : d(z, )


} of . For simplicity we set g = f f .
We claim that

({x} (x)).
(1.4)
{z X X : g n (z) U , n Z} =
xX

In fact, take z = (x, y) in the left-hand side set. Then, for all
n Z there is pn X such that d(f n (x), pn ) + d(f n (y), pn )
so d(f n (x), f n (y)) for all n Z which implies y (x). Therefore z belongs to the right-hand side set. Conversely, if z = (x, y) is
in the right-hand side set then d(f n (x), f n (y)) for all n Z so
d(g n (x, y), (f n (x), f n (x))) = d(f n (x), f n (y)) for all n Z which
implies that z belongs to the left-hand side set. The claim is proved.
Let F be the characteristic map of the left-hand side set in (1.4).
It follows that F (x, y) = (x) (y) for all (x, y) X X where A if
the characteristic map of A X. So,
2 ({z X X : g n (z) U , n Z}) =
 
X

(x) (y)d(y)d(x).

(1.5)

Now suppose that is an expansive measure of f with constant . It


follows that

X

(x) (y)d(y) = 0,

x X

therefore 2 ({z X X : g n (z) U , n Z}) = 0 by (1.5).


Conversely, if 2 ({z X X : g n (z) U , n Z}) = 0 for some
> 0, then (1.5) implies that ( (x)) = 0 for -almost every x X.
Then, is expansive by Lemma 1.20. This ends the proof.
Our nal equivalence is given by using the idea of generators (see
[89]). Call a nite open covering A of X -generator of a homeomorphism f if for every bisequence {An : n Z} A one has



n
f (Cl(An )) = 0.

nZ

15

[SEC. 1.4: PROPERTIES

Theorem 1.24. Let f : X X be a homeomorphism of a compact


metric space X. Then, a Borel probability measure is an expansive
measure of f if and only if f has a -generator.
Proof. First suppose that is expansive and let be its expansivity
constant. Take A as the collection of the open -balls centered at
x X. Then, for any bisequence An A one has


f n (Cl(An )) (x),
x
f n (Cl(An )),
nZ

nZ

so


f n (Cl(An ))

( (x)) = 0.

nZ

Therefore, A is a -generator of f .
Conversely, suppose that f has a -generator A and let > 0 be
a Lebesgue number of A. If x X, then for every n Z there is
An A such that the closed -ball around f n (x) belongs to Cl(An ).
It follows that

f n (Cl(An ))
(x)
nN

so ( (x)) = 0 since A is a -generator.

1.4

Properties

Consider any map f : X X in a metric space X. We already


dened the omega-limit set (z) of z. In the invertible case we also
dene the alpha-limit set


(z) = y X : y = lim f nk (z) for some sequence nk .
k

Following [74] we say that z is a point with converging semiorbits


under a bijective map f : X X if both (z) and (z) reduce to
singleton.
Denote by A(f ) the set of points with converging semiorbits under
f.

16

[CAP. 1: EXPANSIVE MEASURES

An useful tool to study A(f ) is as follows. For all x, y X,


n N+ and m N we dene A(x, y, n, m) as the set of points z X
satisfying


1
i
i
A(x, y, n, m) = z : max{d(f (z), x), d(f (z), y)} , i m .
n
An useful property of this set is given by the following lemma.
Lemma 1.25. For every bijective map f : X X of a separable
metric space X there is a sequence xk X satisfying


A(f )
A(xk , xk , n, m).
(1.6)
nN+ k,k ,mN+

Proof. Since X is separable there is a dense sequence xk . Take z


A(f ) and n N+ . As z A(f ) there are points x, y such that
(z) = x and (z) = y. Then, there is m N+ such that
max{d(f i (z), x), d(f i (z), y)}

1
,
2n

i m.

Since xk is dense there are k, k  N+ such that


max{d(x, xk ), d(y, xk )}

1
.
2n

Therefore,
d(f i (z), xk ) d(f i (z), x) + d(x, xk )

1
1
1
+
= ,
2n 2n
n

and, analogously,
d(f i (z), xk ) d(f i (z), x) + d(x, xk )

1
1
1
+
= ,
2n 2n
n

for all i m proving z A(xk , xk , n, m) so (1.6) holds.


An old result by Reddy [74] is stated below. For completeness we
include its proof here (for another proof see Theorem 2.2.22 in [5]).
Theorem 1.26. The set of points with converging semiorbits under
a expansive homeomorphism of a compact metric space is countable.

17

[SEC. 1.4: PROPERTIES

Proof. Let f : X X be the expansive homeomorphism in the


statement. Since compact metric spaces are separable we can choose
that A(f )
a sequence xk as in Lemma 1.25. Suppose by contradiction

is uncountable. Applying (1.6) we see that
A(xk , xk , n, m)
+

k,k ,mN+

is uncountable for all n N . Fix an expansivity constant e of f and


a positive integer n with n1 2e . Then, there are k, k  , m N such
that A(xk , xk , n, m) is uncountable (and so innite). Therefore, as
X is compact, there are distinct z, w A(xk , xk , n, m) such that
d(f i (z), f i (w)) < e,

|i| m.

As z, w A(xk , xk , n, m) we also have


d(f i (z), f i (w)) d(f i (z), xk ) + d(f i (w), xk )

e e
+ =e
2 2

and
d(f i (z), f i (w)) d(f i (z), xk ) + d(f i (w), xk )
e e
+ = e,
|i| m.
2 2
Consequently w e (z) contradicting that e is an expansivity constant of f . Therefore A(f ) is countable and the proof follows.
In light of this result we can ask if there is a version of it for
expansive measures. Since countable sets corresponds naturally to
zero measure sets it seems natural to prove the following result. Its
proof follows by adapting the aforementioned proof of Theorem 1.26
to the measure theoretical context.
Theorem 1.27. The set of points with converging semiorbits under
a homeomorphism of a separable metric space has zero measure with
respect to any expansive measure.
Proof. Let f : X X the homeomorphism in the statement and xk
be a sequence as in Lemma 1.25. Suppose by contradiction that there
is an expansive measure such that (A(f )) > 0. Applying (1.6) we
get


A(xk , xk , n, m) > 0
n N+ .

k,k ,mN+

18

[CAP. 1: EXPANSIVE MEASURES

Fix an expansivity constant e of and a positive integer n 2e . By


the previous inequality there are k, k  , m N such that
(A(xk , xk , n, m)) > 0.
Let us prove that there is z A(xk , xk , n, m) and 0 > 0 satisfying
(A(xk , xk , n, m) B[z, ]) > 0,

0 < < 0 ,

(1.7)

where B[, ] indicates the closed -ball operation.


Otherwise, for every z A(xk , xk , n, m) we could nd z > 0
such that
(A(xk , xk , n, m) B[z, z ]) = 0.
Clearly




z
B z,
: z A(xk , xk , n, m)
2

is an open covering of A(xk , xk , n, m). As X is a separable metric


space we have that A(xk , xk , n, m) also does, and, since separable
metric spaces are Lindelof, we have that the above open covering has
a countable subcover {Bi : i N} (say). Therefore,

(A(xk , xk , n, m))
(A(xk , xk , n, m) Bi ) = 0
iN

which is absurd. This proves the existence of z and 0 > 0 satisfying


(1.7).
On the other hand, as f is continuous, and both z and m are
xed, we can also nd 0 < 1 < 0 satisfying
d(f i (z), f i (w))

e
whenever |i| m and d(z, w) < 1 .
2

We claim that
A(xk , xk , n, m) B[z, 1 ] e (z).
Indeed, take w A(xk , xk , n, m) B[z, 1 ].
Since w B[z, 1 ] one has d(z, w) < 1 so
d(f i (w), f i (z)) e,

m i m.

19

[SEC. 1.4: PROPERTIES

Since z, w A(xk , xk , n, m) and

1
n

e
2

one has

d(f i (w), f i (z)) d(f i (w), xk ) + d(f i (z), xk ) e


and
d(f i (w), f i (z)) d(f i (w), xk ) + d(f i (z), xk ) e,

|i| m.

All this together yield w e (z) and the claim follows. Therefore,
0 < (A(xk , xk , n, m) B[z, 1 ]) (e (z))
which is absurd since e is an expansivity constant. This ends the
proof.
Remark 1.28. If f is an expansive homeomorphism of a compact
metric space, then every non-atomic Borel probability measure is an
expansive measure of f . Then, Theorem 1.27 implies that the set
of points with converging semiorbits under f has zero measure with
respect to any non-atomic Borel probability measure. From this and
well-known measure-theoretical results [73] we obtain that the set of
points with converging semiorbits under f is countable. This provides
another proof of the Reddys result [74].
The following lemma will be useful in the next proof.
Lemma 1.29 (see Lemma 4 p. 72 in [16]). If f : X X is a
continuous map of a compact metric space X and (x) is nite for
some x X, then there is a periodic point z X of f such that
d(f n (x), f n (z)) 0 as n .
Proof. Take any nonempty proper closed subset F (x). We claim
that F Cl((x) \ F )
= . Otherwise there are open sets O1 , O2 such
that (x)\F O1 , F O2 and Cl(O2 )f (Cl(O1 )) = . For n large,
f n (x) belongs to O1 or O2 and in both for innitely many n s. Then,
there is an innite sequence nk with f nk (x) O1 and f nk +1 (x) O2 .
Any limit point y of f nk (x) satises y Cl(O1 ) f 1 (Cl(O2 )) thus
Cl(O2 ) f (Cl(O1 ))
= which is absurd. This proves the claim.
Since (x) is nite there is a periodic orbit P (x). If P
= (x)
we could apply the claim to the closed subset F = (x) \ P yielding
((x) \ P ) P
= which is absurd. Therefore, P = (x) from which
the result easily follows.

20

[CAP. 1: EXPANSIVE MEASURES

By heteroclinic point of a bijective map f : X X on a metric space X we mean any point for which both the alpha and the
omega-limit sets reduce to periodic orbits. The lemma below relates
homoclinic and points with converging semiorbits. Denote by Het(f )
the set of heteroclinic points of f .
Lemma 1.30. If f : X X is a homeomorphism of a compact
metric space X, then

A(f n ).
Het(f )
nN+

Proof. If x Het(f ), then both (x) and (x) are nite sets. Applying Lemma 1.29 we get a periodic point y such that d(f n (x), f n (y))
0 as n . Denoting by ny the period of y we get d(f kny (x), y) 0
as k and so f ny (x) = {y}. Analogously, f nz (x) = {z} for
some periodic point z of period nz . Taking n = ny nz we obtain
n N+ such that f n (x) = z and f n (x) = y so x A(f n ) and the
inclusion follows.
Theorem 1.27 and Lemma 1.30 have the following consequence.
Theorem 1.31. The set of heteroclinic points of a homemorphism in
a compact metric space has measure zero with respect to any expansive
measure.
Proof. Let f : X X be a homeomorphism of a compact metric
space. By Lemma 1.30 we have
that the set of heteroclinic points
A(f n ). Now, take any expansive
satises the inclusion Het(f )
nN+

measure of f . By Lemma 1.22 we have that is also an expansive


measure of f n , and so, (A(f n )) = 0 for all n N+ by Theorem
1.27. Then, the inclusion above implies

(A(f n )) = 0
(Het(f ))
nN+

proving the result.


A consequence of the above result is given below.

[SEC. 1.4: PROPERTIES

21

Corollary 1.32. A homeomorphism with nite nonwandering set of


a compact metric space has no expansive measures.
Proof. This follows from Theorem 1.31 since every point for such
homeomorphisms is heteroclinic.
(In the probability case this corollary is a particular case of Corollary 1.19).
Another consequence is the following version of Theorem 3.1 in
[86]. Denote by Per(f ) the set of periodic points of f .
Corollary 1.33. The set of periodic points of a homeomorphism of a
compact metric space has measure zero with respect to any expansive
measure.
Proof. Let be an expansive measure of a homeomorphism f of a
compact metric space. Denoting by Fix(f ) = {x X : f (x) = x} the
set of xed points of a map f we have Per(f ) = nN+ Fix(f n ). Now,
is an expansive measure of f n by Proposition 1.22 and every element
n
of Fix(f n ) is a heteroclinic point of f n thus
 (Fix(f )) =n 0 for all n by
Theorem 1.27. Therefore, (Per(f )) nN+ (Fix(f )) = 0.
We nish this section by describing the expansive measures in dimension one. To start with we prove that there are no such measures
for homeomorphisms of compact intervals.
Theorem 1.34. A homeomorphism of a compact interval has no
expansive measures.
Proof. Suppose by contradiction that there is an expansive measure
for some homeomorphism f of I. Since f is continuous we have
that Fix(f )
= . Such a set is also closed since f is continuous,
so, its complement I\Fix(f ) in I consists of countably many open
intervals J. It is also clear that every point in J is a point with
converging semi-orbits therefore (I\ Fix(f )) = 0 by Theorem 1.27.
But (Fix(f )) = 0 by Corollary 1.33 so (I) = (Fix(f )) + (I\
Fix(f )) = 0 which is absurd.
Next, we shall consider the circle S 1 . Recall that an orientationpreserving homeomorphism of the circle S 1 is Denjoy if it is not
topologically conjugated to a rotation [40].

22

[CAP. 1: EXPANSIVE MEASURES

Theorem 1.35. A circle homeomorphism has expansive measures if


and only if it is Denjoy.
Proof. Let f be a Denjoy homeomorphism of S 1 . As is well known
f has no periodic points and exhibits a unique minimal set which
is a Cantor set [40]. In particular, is compact without isolated
points thus it exhibits a non-atomic Borel probability meeasure (c.f.
Corollary 6.1 in [73]). On the other hand, one sees as in Example 1.2
of [25] that f / is expansive so is an expansive measure of f /.
Then, we are done by Example 1.12.
Conversely, let be an expansive measure of a homeomorphism
f : S 1 S 1 and suppose by contradiction that f is not Denjoy. Then,
either f has periodic points or is conjugated to a rotation (c.f. [40]).
In the rst case we can assume by Proposition 1.22 that f has a xed
point. Then, we can cut open S 1 along the xed point to obtain an
expansive measure for some homeomorphism of I which contradicts
Theorem 1.34. In the second case we have that f is conjugated to
a rotation. Since is expansive it would follow from Example 1.13
that there are circle rotations with expansive measures. However,
such rotations cannot exist by Example 1.9 since they are isometries.
This contradiction proves the result.
In particular, there are no expansive measures for C 2 dieomorphisms of S 1 . Similarly, there are no such measures for dieomorphisms of S 1 with derivative of bounded variation.

1.5

Probabilistic proofs in expansive systems

The goal of this short section is to present the proof of some results
in expansive systems using the ours.
To start with we shall prove the following result.
Proposition 1.36. The set of periodic points of a measure-expansive
homeomorphism f : X X of a compact metric space X is countable.
Proof. Since Per(f ) = nN+ Fix(f n ) it suces to prove that Fix(f n )
is countable for all n N+ . Suppose by contradiction that Fix(f n ) is

[SEC. 1.5: PROBABILISTIC PROOFS IN EXPANSIVE SYSTEMS

23

uncountable for some n. Since f is continuous we have that Fix(f n )


is also closed, so, it is complete and separable with respect to the
induced topology. Thus, by Corollary 6.1 p. 210 in [73], there
is a non-atomic Borel probability measure in Fix(f n ). Taking
(A) = (Y A) for all borelian A of X we obtain a non-atomic
Borel probability measure of X satisfying (Fix(f n )) = 1. Since
Fix(f n ) Per(f ) we conclude that (Per(f )) = 1. However, is
an expansive measure of f thus (Per(f )) = 0 by Corollary 1.33, a
contradiction. This contradiction yields the result.
Since every expansive homeomorphism of a compact metric space
is measure-expansive the above proposition yields another proof of
the following result due to Utz (see Theorem 3.1 in [86]).
Corollary 1.37. The set of periodic points of an expansive homeomorphism of a compact metric space is countable.
A second result is as follows.
Proposition 1.38. Measure-expansive homeomorphisms of compact
intervals do not exist.
Proof. Suppose by contradiction that there is a measure-expansive
homeomorphism of a compact interval I. Since the Lebesgue measure
Leb of I is non-atomic we obtain that Leb is an expansive measure
of f . However, there are no such measures for such homeomorphisms
by Theorem 1.34.
From this we obtain another proof of the following result by Jacobson and Utz [46] (details in [19]).
Corollary 1.39. There are no expansive homeomorphisms of a compact interval.
The following lemma is motivated by the well known property
that for every homeomorphism f of a compact metric space X one
has that supp() (f ) for all invariant Borel probability measure
of f . Indeed, we shall prove that this is true also for all expansive
measure oif every homeomorphism of S 1 even in the noninvariant
case.

24

[CAP. 1: EXPANSIVE MEASURES

Lemma 1.40. If f : S 1 S 1 is a homeomorphism, then supp()


(f ) for every expansive measure of f .
Proof. Suppose by contradiction that there is x supp() \ (f ) for
some expansive measure of f . Let be an expansivity constant of
. Since x
/ (f ) we can assume that the collection of open intervals
f n (B(x, )) as n runs over Z is disjoint. Therefore, there is N N
such that the length of f n (B(x, )) is less than for |n| N .
From this and the continuity of f we can arrange  > 0 such that
B(x, ) (x) therefore ( (x)) (B(x, )) > 0 as x supp().
This contradicts the expansiveness of and the result follows.
A direct consequence of this lemma is the following.
Corollary 1.41. A homeomorphism of S 1 has no expansive measures
supported on S 1 .
Proof. Suppose by contradiction that there is a homeomorphism f :
S 1 S 1 exhibiting an expansive measure with supp() = S 1 . By
Theorem 1.35 we have that f is Denjoy, and so, (f ) is nowhere
dense. However, we have by Lemma 1.40 that supp() (f ) so S 1
is nowhere dense too which is absurd.
This corollary implies immediately the following one.
Corollary 1.42. There are no measure-expansive homeomorphisms
of S 1 .
Proof. If there were such homeomorphisms in S 1 , then the Lebesgue
measure would be an expansive measure of some homeomorphism of
S 1 contradicting Corollary 1.41.
From this we obtain the following classical fact due to Jacobsen
and Utz [46]. Classical proofs can be found in Theorem 2.2.26 in [5],
Subsection 2.2 of [25], Corollary 2 in [74] and Theorem 5.27 of [89].
Corollary 1.43. There are no expansive homeomorphisms of S 1 .

25

[SEC. 1.6: EXERCICES

1.6

Exercices

Exercice 1.44.

Prove that the set of heteroclinic points of a homeomorphism


of a compact metric space is Borel measurable.

Exercice 1.45.

Are there homeomorphisms of compact metric spaces exhibiting

a unique expansive measure? (Just in case prove that such a measure is invariant).

Exercice 1.46.

Are there measure-expansive homeomorphisms of compact nonatomic metric spaces which are not countably-expansive?

Exercice 1.47.

Prove (or disprove) that every homeomorphism possessing an


expansive probability measure on a compact metric space also possesses ergodic expansive invariant probability measures.

Exercice 1.48.

Are there measure-expansive homeomorphisms of S 2 ?

Exercice 1.49.

It is well known that, for expansive homeomorphisms f on compact metric spaces, the entropy map  h (f ) is uppersemicontinuous [89]. Are
the expansive measures for homeomorphisms on compact metric space uppersemicontinuity points of the corresponding entropy map?

Exercice 1.50.

It is well known that every compact metric space supporting


expansive homeomorphisms has nite topological dimension [62]. Is the support
of an expansive measure of a homeomorphisms of a compact metric space nite
dimensional?

Exercice 1.51.

A bijective map f : X X of a metric space X is distal if


inf d(f n (x), f n (y)) > 0,

nZ

x X.

It is well known that a distal homeomorphism has zero topological entropy ([8],[36],
[68]). Are there distal homeomorphisms with expansive measures of compact metric
spaces.

Exercice 1.52.

A generalization of the previows problem can be stated as fol-

lows. A Li-Yorke pair of a continuous map f : X X is a pair (x, y) X X


which is proximal (i.e. lim inf n d(f n (x), f n (y)) = 0) but not asymptotic (i.e.
lim supn d(f n (x), f n (y)) > 0). We say that f is almost distal if it has no LiYorke pairs. Every distal homeomorphism is almost distal but not conversely. On the
other hand, almost distal maps on compact metric space have some similarities with
the distal ones as, in particular, all of them have zero topological entropy [15]. Prove
(or disprove) that every almost distal homeomorphism of a compact metric space has
expansive measures.

26

[CAP. 1: EXPANSIVE MEASURES

Exercice 1.53.

Prove that the space of expansive measures Mexp (f ) of a measurable map f : X X on a metric space X is a cone, i.e., + Mexp (f )
whenever R+ and , Mexp (f ). Furthermore, if : X Y is a conjugation
between f and another measurable map g : Y Y of a metric space Y , then
f (Mexp (f )) = Mexp (g).

Exercice 1.54.

Prove that if f : S 1 S 1 is a local homeomorphism of the


circle S 1 , then the Lebesgue measure is expansive for f if and only if f is expansive.

Exercice 1.55.

Given metric spaces (X, dX ) and (Y, dY ) we dene the metric


dXY ((x1 , y1 ), (x2 , y2 )) = max{dX (x1 , y1 ), dY (x2 , y2 )} in X Y . With respect

to this metric prove that if and are expansive measures of the homeomorphisms
f : X X and g : Y Y , then so is the product measure of X Y for the
product map f g. Give a counterexample for the converse (see Exercice 1.56).

Exercice 1.56.

Let f : X X be a homeomorphism of a metric space X.

Prove that a Borel measure of X is expansive for f if and only if the product
measure Leb of with the Lebesgue measure Leb of [0, 1] is expansive for the
product map f Id : X [0, 1] X [0, 1], where Id is the identity map of [0, 1].

Exercice 1.57.

Prove that a homeomorphim f : D D of the closed unit

2-disk D R2 for which the alpha-limit set (x) = (0, 0) for all x Int(D) has no
expansive measures.

Exercice 1.58.

We say that a homeomorphism f : X X of a metric space


X is proximal if inf nZ d(f n (x), f n (y)) = 0 for every x, y X. Find examples

of proximal homeomorphisms of compact metric spaces with and without expansive


measures.

Exercice 1.59.

Motivated by [75] we call a non-trivial Borel measure of a


metric space X pointwise expansive for a homeomorphism f : X X if for every
x X there is x > 0 such that (x (x)) = 0. Investigate the vality (or not) of the
results of this chapter for pointwise expansive measures instead of expansive ones.

Exercice 1.60. Prove that the property of being an expansive measure is a


metric invariant in the following sense: If f : X X is a homeomorphism of a
metric space (X, d) and d is a metric of X equivalent to d, then a Borel measure is
expansive for f if and only if it does for f : (X, d ) (X, d ).

Chapter 2

Finite expansivity
2.1

Introduction

We already seem that every expansive homeomorphism of a nonatomic metric space is measure-expansive (i.e. it satises that every
non-atomic Borel probability measure is an expansive measure). It is
then natural to ask if the converse property holds, i.e., is a measureexpansive homeomorphism of a non-atomic metric space expansive?
The results of this chapter will provide negative answer for this question even on compact metric spaces (see Exercice 3.44).

2.2

Preliminaries

In this section we establish some topological preliminaries. Let X a


set and n be a nonnegative integer. Denote by #A the cardinality of
A. The set of metrics of X (including -metrics [32]) will be denoted
by M(X). Sometimes we say that M(X) has a certain property
whenever its underlying metric space (X, ) does. For example, is
compact whenever (X, ) is, a point a is -isolated in A X if it is
isolated in A with respect to the metric space (X, ), etc.. The closure
operation in (X, ) will be denoted by Cl (). A map f : X X
is a -homeomorphism if it is a homeomorphism of the metric space
(X, ). If x X and > 0 we denote by B [x, ] the closed -ball
27

28

[CAP. 2: FINITE EXPANSIVITY

around x (or B[x, ] if there is no confusion).


Given M(X) and A X we say that is n-discrete on A if
there is > 0 such that #(B[x, ]A) n for all x A. Equivalently,
if there is > 0 such that #(B[x, ] A) n for all x X. When
necessary we emphasize by saying that is n-discrete on A with
constant . We say that is n-discrete if it is n-discrete on X. Clearly
is n-discrete on A if and only if the restricted metric /A M(A)
dened by /A(a, b) = (a, b) for a, b A is n-discrete.
Evidently, there are no 0-discrete metrics and the 1-discrete metrics are precisely the discrete ones. Since every n-discrete metric is
m-discrete for n m one has that every discrete metric is n-discrete.
There are however n-discrete metrics which are not discrete. Moreover, we have the following example (1 ).
Example 2.1. Every innite set X carries an n-discrete metric
which is not (n 1)-discrete.
Indeed, if n = 1 we simply choose as the standard discrete
metric (x, y) dened by (x, y) = 1 whenever x
= y. Otherwise, we
can arrange n disjoint sequences x1k , x2k , xnk in X and dene by
1
(if (x, y) = (xik , xjk ) for some k N and 1 i
= j n)
(x, y) = 4+k
and (x, y) = (x, y) (if not).


On the one hand, is n-discrete with constant = 14 since B x, 14
is either {x1k , , xnk } or {x} (depending on the case) and, on the
other, is not (n 1)-discrete since for all > 0 the set of points x
for which #B[x, ] = n is innite (e.g. take x = x1k with k large).
Remark 2.2. None of the metrics in Example 2.1 can be compact
for, otherwise, we could cover X with nitely many balls of radius
= 1/4 which would imply that X is nite.
In the sequel we present some basic properties of n-discrete metrics. Clearly if is n-discrete on A, then it is also n-discrete on B for
all B A. Moreover, if is n-discrete on A and m-discrete on B,
then it is (n + m)-discrete on A B. A better conclusion is obtained
when the distance between A and B is positive.
Lemma 2.3. If is n-discrete on A, m-discrete on B and (A, B) >
0, then is max{n, m}-discrete on A B.
1 communicated

by professors L. Florit and A. Iusem.

[SEC. 2.2: PRELIMINARIES

29

Proof. Choose 0 < < (A,B)


such that #(B[x, ] A) n (for x
2
A) and #(B[x, ] B) m (for x B). If x A then B[x, ] B =
so #(B[x, ] (A B)) = #(B[x, ] A) n
because < (A,B)
2
max{n, m}. If x B then B[x, ] A = because < (A,B)
so
2
#(B[x, ] (A B)) = #(B[x, ] B) m max{n, m}. Then, is
max{n, m}-discrete on A B with constant .
Lemma 2.4. If is n-discrete on A, then A is -closed and so
(A, B) > 0 for every -compact subset B with A B = .
Proof. We only have to prove the rst part of the lemma. By hypothesis there is > 0 such that #(B[x, ] A) n for all x A. Let xk
be a sequence in A converging to some y X. It follows that there is
k0 N+ such that xk B[y, /2] for all k k0 . Triangle inequality
implies {xk : k k0 } B[xk0 , ] A and so {xk : k k0 } is a nite
set. As xk y we conclude that y A hence A is closed.
Now we prove that n-discreteness is preserved under addition of
nite subsets.
Proposition 2.5. If is n-discrete on A, then is n-discrete on
A F for all nite F X.
Proof. We can assume that AF = . As F is nite (hence compact)
we can apply Lemma 2.4 to obtain (A, F ) > 0. As F is nite one
has that is 1-discrete on F so is n-discrete on A F by Lemma
2.3.
For the next result we introduce some basic denitions. Let f :
X X be a map. We say that A X is invariant if f (A) = A. If
f is bijective and x X we denote by Of (x) = {f n (x) : n Z} the
orbit of x. An isometry (or -isometry to emphasize ) is a bijective
map f satisfying (f (x), f (y)) = (x, y) for all x, y X.
The following elementary fact will be useful later one: If f is a
-isometry and a X satises that a is -isolated in Of (a), then is
discrete on Of (a). Indeed, if were not discrete on Of (a), then there
are integer sequences nk
= mk such that (f nk (a), f mk ) 0 as k
. As f is an isometry one has that (f nk (a), f mk (a)) = (a, f lk (a)),
where lk = mk mk , so (a, f lk (a)) 0 for some sequence lk Z\{0}
thus a is not -isolated in Of (a).

30

[CAP. 2: FINITE EXPANSIVITY

Given d, M(X) we write d whenever d(x, y) (x, y) for


all x, y X. We write  d to indicate lower semicontinuity of the
map : X X [0, ] with respect to the product metric d d in
X X. Equivalently, the following property holds for all sequences
d
xk , yk in X and all > 0, where xk x indicates convergence in
(X, d):
d

xk x,

yk y

y B [x, ].
(2.1)
Hereafter we denote by F ix(f ) = {x X : f (x) = x} the set
of xed points of f , and by P er(f ) = mN+ F ix(f m ) the set of
periodic points of f .
The following proposition is inspired on Lemma 2 p. 176 of [89].
and

yk B [xk , ]

Proposition 2.6. Let d, M(X) be such that d is compact and


d  d. Let f : X X be a map which is simultaneously
a d-homeomorphism and a -isometry. If A is an invariant set
with countable complement which is n-discrete with respect to and
P er(f ) A is countable, then is n-discrete on A Of (a) for all
a X.
Proof. We can assume a
A (otherwise A Of (a) = A) so (A, a) >
0 by Lemma 2.4. Since f is a -isometry and A is invariant one has
(A, f i (a)) = (A, a) so (A, Of (a)) > 0. Then, by Lemma 2.3, it
suces to prove that is n-discrete on Of (a).
Suppose that it is not so. Then, as previously remarked, a is non
-isolated in Of (a). Since d we have that a is also non -isolated
in Of (a). As f is a d-homeomorphism we conclude that Of (a) is a
nonempty -perfect set. As d is compact (and so FII ) we obtain that
Cld (Of (a)) is uncountable. As X \ A is countable we conclude that
Cld (Of (a)) A is uncountable. Choose x Cld (Of (a)) A. Then,
there is a sequence lk Z such that
d

f lk (a) x.

(2.2)

Let > 0 be such that is n-discrete on A with constant .


Since is not n-discrete on Of (a) we can arrange dierent integers
N1 , , Nn+1 satisfying
f Nj (a) B [f N1 (a), ],

j {1, , n + 1}.

31

[SEC. 2.2: PRELIMINARIES

On the other hand, f is a -isometry so the above inclusions yield


f Nj (f lk (a)) B [f N1 (f lk (a)), ],

j {1, , n + 1},

k N.

By taking limit as k in the above inclusion, keeping j xed and


applying (2.1) and (2.2) to obtain
f Nj (x) B [f N1 (x), ],

j {1, , n + 1}.

Now observe that f Nj (x) A for all j {1, , n + 1} because A is


invariant. Therefore,
{f N1 (x), , f Nn+1 (x)} B [f N1 (x), ] A.
But #(B [f N1 (x), ]A) n by the choice of so the above inclusion
implies f Nj (x) = f Nr (x) for some dierent indexes j, r {1, , n +
1}. As the integers N1 , , Nn+1 are dierent we conclute that x
P er(f ) and so x P er(f ) A. Therefore,
Cld (Of (a)) A P er(f ) A.
As Cld (Of (a)) A is uncountable we conclude that P er(f ) A also
is thus we get a contradiction. This proves the result.
Corollary 2.7. Let d, M(X) be such that d is compact and d
 d. Let f : X X be a map which is simultaneously a dhomeomorphism and a -isometry. If P er(f ) is countable and there
l
are a1 , , al X such that is n-discrete on X \ i=1 Of (ai ), then
is n-discrete.
l
Proof. Dene the invariant sets Aj = X \ i=j Of (ai ) for 1 j l.
l
As X \ Aj = i=j Of (ai ) one has that Aj has countable complement
for all 1 j l. On the other hand, is n-discrete on A1 by
hypothesis and P er(f ) A1 is countable (since P er(f ) is) so is
n-discrete on A2 = A1 Of (a1 ) by Proposition 2.6. By the same
reasons if is n-discrete on Aj , then also is on Aj+1 = Aj Of (ai ).
Then, the result follows by induccion.

32

2.3

[CAP. 2: FINITE EXPANSIVITY

n-expansive systems

In this section we dene and study the class of n-expansive systems.


To motivate the denition we recall some classical denitions. Let
(X, d) be a metric space and A X. A map f : X X is positively
expansive on A if there is > 0 such that for every x, y A with
x
= y there is i N such that d(f i (x), f i (y)) > , or, equivalently,
if {y A : d(f i (x), f i (y)) , i N} = {x} for all x A. On the
other hand, a bijective map f : X X is expansive on A if there is
> 0 such that {y A : d(f i (x), f i (y)) , i Z} = {x} for all
x A. If A = X we recover the notions of positively expansive and
expansive maps respectively. These denitions suggest the following
one.
Denition 2.8. Given n N+ a bijective map (resp. map) f is nexpansive (resp. positively n-expansive) on A if there is > 0 such
that
#{y A : d(f i (x), f i (y)) , i Z} n
(resp. #{y A : d(f i (x), f i (y)) , i N} n)
x A. If case A = X we say that f is n-expansive (resp. positively
n-expansive).
Clearly the 1-expansive bijective maps are precisely the expansive
ones (which in turn are n-expansive for all n N+ ).
In the sequel we introduce two useful operators. For every f :
X X and d M(X) we dene the pull-back metric
f (d)(x, y) = d(f (x), f (y))
(clearly f (d) M(X) if and only if f is 1-1). Using it we can dene
the operator L+
f : M(X) M(X) by
i
L+
f (d) = sup f (d),
iN

d M(X).

If f is bijective we can dene Lf : M(X) M(X) by


Lf (d) = sup fi (d),
iZ

d M(X).

33

[SEC. 2.3: N -EXPANSIVE SYSTEMS

Lemma 2.9. If f is bijective, then d Lf (d) and f is a Lf (d)isometry. If, in addition, f is a d-homeomorphism, then Lf (d)  d.
Proof. The rst inequality is evident. As
f (Lf (d))(x, y) =
sup d(f i+1 (x), f i+1 (y)) = sup d(f i (x), f i (y)) = Lf (d)(x, y)
iZ

iZ

(x, y X) one has f (Lf (d)) = Lf (d) hence f is an Lf (d)-isometry.


Now we prove Lf (d)  d whenever f is a d-homeomorphism. Suppose
d

that xk x, yk y and Lf (d)(xk , yk ) for all k N. Fixing


i Z the latter inequality implies d(f i (xk ), f i (yk )) for all k. As
f is a d-homeomorphism one can take the limit as k in the last
inequality to obtain d(f i (x), f i (y)) . As i Z is arbitrary we
obtain Lf (d)(x, y) which together with (2.2) implies the result.
These operators give the link between discreteness and expansiveness by the following result. Hereafter we shall write f is (positively)
n-expansive (on A) with respect to d in order to emphazise the metric
d in Denition 2.8.
Lemma 2.10. The following properties hold for all f : X X,
A X and d M(X):
1. f is positively n-expansive on A with respect to d if and only if
L+
f (d) is n-discrete on A.
2. If f is bijective, f is n-expansive on A with respect to d if and
only if Lf (d) is n-discrete on A.
Proof. Clearly for all x X and > 0 one has
+

B Lf (d) [x, ] A = {y A : d(f i (x), f i (y)) ,


so

#(B Lf (d) [x, ] A) n


#({y A : d(f (x), f (y)) ,
i

i N},

i N}) n

which proves the equivalence (1). The proof of the equivalence (2) is
analogous.

34

[CAP. 2: FINITE EXPANSIVITY

As a rst application of the above equivalence we shall exhibit


non-trivial examples of positively n-expansive maps. More precisely,
we prove that every bijective map f : X X with at least n nonperiodic points (n 2) carries a metric making it continuous positively n-expansive but not positively (n 1)-expansive. Indeed, by
hypothesis there are x1 , , xn X such that f i (xj )
= f k (xj ), for
all 1 j n and i
= k N, and f i (xj )
= f i (xk ) for all i N and
1 j
= k n. Dene the sequences x1k , , xnk in X by xik = f k (xi )
for 1 i n and k N. Clearly these sequences are disjoint thus
they induce a metric in X which is n-discrete but not (n 1)discrete as in Example 2.1. On the other hand, a straightforward
computation yields L+
f () = thus f is continuous (in fact Lips+
chitz) for . Since is n-discrete and = L+
f () one has that Lf ()
is n-discrete so f is positively n-expansive by Lemma 2.10. Since
is not (n 1)-discrete and = L+
f () the same lemma implies that
f is not positively (n 1)-expansive.
Notice however that none of the above metrics is compact (see
for instance Remark 2.2). This fact leads the question as to whether
a bijective map can carry a compact metric making it positively nexpansive but not positively (n 1)-expansive. Indeed, the following
result gives a partial positive answer for this question.
Proposition 2.11. For every k N+ there is a homeomorphism fk
of a compact metric space (Xk , k ) which is positively 2k -expansive
but not positively (2k 1)-expansive.
Proof. To start with we recall that a Denjoy map of the circle S 1 is a
nontransitive homeomorphism of S 1 with irrational rotation number.
As is well known [40] every Denjoy map h exhibits a unique minimal
set Eh which is also a Cantor set.
Hereafter we x the standard Riemannian metric l of S 1 . We shall
prove that h/Eh is positively 2-expansive with respect to l/Eh . Let
be half of the length of the largest interval I in the complement
S 1 \ Eh and 0 < < .
+
We claim that Int(B Lh (l) [x, ]) Eh = for all x Eh . Oth+
erwise, there is some z Int(B Lh (l) [x, ]) Eh . Pick w I (thus
w Eh ). Since Eh is minimal there is a sequence nk such
that hnk (w) z. Now, the interval sequence {hn (I) : n N}

35

[SEC. 2.3: N -EXPANSIVE SYSTEMS

is disjoint so we have that the length of the intervals hnk (I) 0


as k . It turns out that there is some integer k such that
+
+
hnk (I) B Lh (l) [x, ]. From this and the fact that h(B Lh (l) [x, ])
+
+
B Lh (l) [h(x), ] one sees that I B Lh (l) [hnk (x), ] which is clearly
absurd because the length of I is greather than > 2. This contradiction proves the claim.
+
Since B Lh (l) [x, ] reduces to closed interval (possibly trivial) the
+
claim implies that B Lh (l) [x, ] Eh consists of at most two points.
It follows that L+
h (l) is 2-discrete on Eh (with constant ), so, h/Eh
is positively 2-expansive with respect to l/Eh by Lemma 2.10. Since
there are no positively expansive homeomorphisms on innite compact metric spaces (e.g. [28]) one sees that h/Eh cannot be positively
expansive with respect to l/Eh . Taking X1 = Eh , 1 = l/Eh and
f1 = h/Eh we obtain the result for k = 1. To obtain the result for
k 2 we shall proceed according to the following straightforward
construction.
Take copies E1 , E2 of Eh and recall the map
max{, } : M(E1 ) M(E2 ) M(E1 E2 )
dened by
max{d1 , d2 }(x, y) = max{d1 (x1 , y1 ), d2 (x2 , y2 )}
for all x = (x1 , x2 ) and y = (y1 , y2 ) in E1 E2 . One clearly sees that
B max{d1 ,d2 } [x, ] = B d1 [x1 , ] B d2 [x2 , ],

x E1 E2 , > 0.

Afterward, take copies h1 , h2 of h/Eh and dene the product h1


h2 : E1 E2 E1 E2 , (h1 h2 )(x) = (h1 (x1 ), h2 (x2 )). It turns
out that
(h1 h2 ) (max{d1 , d2 }) = max{h1 (d1 ), h2 (d2 )}
so

+
+
L+
h1 h2 (max{d1 , d2 }) = max{Lh1 (d1 ), Lh2 (d2 )}

thus
+

B Lh1 h2 (max{d1 ,d2 }) [x, ] = B Lh1 (d1 ) [x1 , ] B Lh2 (d2 ) [x2 , ].

36

[CAP. 2: FINITE EXPANSIVITY

Finally, take copies d1 , d2 of the metric l/Eh each one in E1 , E2


respectively. As hi is positively 2-expansive with respect to di one
has that L+
hi (di ) is 2-discrete for i = 1, 2. We can choose the same
constant for i = 1, 2 ( say) thus,
+

#(B Lh1 h2 (max{d1 ,d2 }) [x, ]) =


#(B

L+
h (d1 )
1

[x1 , ]) #(B

L+
h (d2 )
2

[x2 , ]) 22 ,

x E1 E2 . (2.3)

Now, consider the compact metric space (E1 E2 , max{d1 , d2 }).


It follows from (2.3) and Lemma 2.10 that h1 h2 (which is clearly
a homeomorphism) is positively 22 -expansive map with respect to
L+

(max{d ,d })

1 2
max{d1 , d2 }. One can see that #(B h1 h2
[x, ]) = 22 for
innitely many xs and arbitrarily small thus h1 h2 cannot be
positively 22 1-expansive. Taking X2 = E1 E2 , 2 = max{d1 , d2 }
and f2 = h1 h2 we obtain the result for k = 2.
By repeating this argument we obtain the result for arbitrary
k N+ taking X2 = E1 Ek , k = max{d1 , , dk } and
fk = h1 hk .

As a second application of the equivalence in Lemma 2.10 we


establish the following lemma which is well-known among expansive
systems (e.g. Lemma 1 in [89]).
Lemma 2.12. If a homeomorphism f of a metric space (X, d) is
n-expansive on A, then P er(f ) A is countable.
Proof. It follows from the hypothesis and Lemma 2.10 that there is
> 0 such that #(B Lf (d) [x, ] A) n for all x X.
First we prove that f m is n-expansive on A, m N+ . Observe
that f is continuous since d is compact so there is  > 0 such that
d(x, y)  implies d(f i (x), f i (y)) for all integer m i m.
Then, B Lf m (d) [x, ] B Lf (d) [x, ] for all x X, so, #(B Lf m (d) [x, ]
A) #(B Lf (d) [x, ] A) n for all x A. Therefore, Lf m (d) is ndiscrete on A (with constant ) which implies that f m is n-expansive
on A by Lemma 2.10.This proves the assertion.
Since P er(f ) = mN+ F ix(f m ) by the previous assertion we
only have to prove that F ix(f )A is nite whenever f is n-expansive
on A. To prove it suppose that there is an innite sequence of xed

37

[SEC. 2.4: THE RESULTS


d

points xk A. Since d is compact one can assume that xn x


for some x X. On the other hand, one clearly has Lf (d) = d in
F ix(f ) thus, by the triangle inequality on x, there is n0 N such that
xn B Lf (d) [xn0 , ] for all n n0 . Thus, #(B Lf (d) [xn0 , ] A) =
which contradicts the choice of above. This ends the proof.

2.4

The results

In this section we state and prove our main results. The rst one
establishes that there are arbitrarily large values of n for which there
are innite compact metric spaces carrying positively n-expansive
homeomorphisms. As is well known, this is not true in the positively
expansive case (see for instance [28]).
Theorem 2.13. For every k N+ there is an innite compact metric space (Xk , k ) carrying positively 2k -expansive homeomorphisms
which are not positively (2k 1)-expansive.
Proof. Take Xk , k and fk as in Proposition 2.11. As fk is not positively (2k 1)-expansive one has that Xk is innite.
From this we obtain the following corollary.
Corollary 2.14. There are compact metric spaces without isolated
points exhibiting homeomorphism which are not positively expansive
but for which every non-atomic Borel probability measure is positively
expansive.
Our second result generalizes the one in [12].
Theorem 2.15. A map (resp. bijective map) of a metric space (X, d)
is positively n-expansive (resp. n-expansive) if and only if it is positively n-expansive (resp. n-expansive) on X \ F for some nite subset
F.
Proof. Obviously we only have to prove the if part. We do it in
the positively n-expansive case as the n-expansive case follows analogously. Suppose that a map f of X is positively n-expansive on X \F
for some nite subset F . Then, L+
f (d) is n-discrete on A = X \ F by
Lemma 2.10. Since F is nite Proposition 2.5 implies that L+
f (d) is
n-discrete so f is positively n-expansive by Lemma 2.10.

38

[CAP. 2: FINITE EXPANSIVITY

Finally we state our last result which extends a well-known property of expansive homeomorphisms (c.f. [85],[89]).
Theorem 2.16. A necessary and sucient condition for a homemomorphism f of a compact metric space (X, d) to be n-expansive is
l
that f is n-expansive on X \ i=1 Of (ai ) for some a1 , , al X.
Proof. We only have to prove the if part. By hypothesis f is a dhomeomorphism so f is an Lf (d)-isometry and d Lf (d)  d by
l
Lemma 2.9. Since f is n-expansive on A = X \ i=1 Of (ai ) one has
l
that P er(f )A is countable by Lemma 2.12. As X \A = i=1 Of (ai )
is clearly countable we conclude that P er(f ) is countable. On the
l
other hand, f is n-expansive on X \ i=1 Of (ai ) so Lf (d) is n-discrete
l
on X \ i=1 Of (ai ) by Lemma 2.10. Then, Lf (d) is n-discrete by
Corollary 2.7 and so f is n-expansive by Lemma 2.10.

2.5

Exercices

Exercice 2.17.
Exercice 2.18.

Prove the assertion in Remark 2.2.


Prove that for every integer n 2 there is an n-expansive

homeomorphism of a compact metric space which is not (n 1)-expansive.

Exercice 2.19.

Prove that every compact metric space with n-expansive homeomorphisms (for some n N+ ) has nite topological dimension and that the minimal
sets of such a homeomorphism are zero-dimensional (for n = 1 see Ma
n
e [62]).

Exercice 2.20. Prove that every n-expansive homeomorphism f : X X of a


metric space is pointwise expansive, i.e., for every x X there is x > 0 such that
x (x) = {x} (see [75]).
Exercice 2.21.

Prove that the non-expansive pointwise expansive homeomor-

phisms dened by Reddy in Section 3 of [75] are 2-expansive. Modify these examples
to nd for all n 2 an n-expansive homeomorphism of a compact metric space which
is not pointwise expansive.

Exercice 2.22.

Are there dierentiable manifolds supporting n-expansive homeomorphisms which are not (n 1)-expansive?

[SEC. 2.5: EXERCICES

Exercice 2.23.

39

Prove that an n-expansive homeomorphism f : X X of a

compact metric space X has measures of maximal entropy, i.e., a Borel measure
satisfying the identity h (f ) = h(f ) where h (f ) and h(f ) denotes the metric and
topological entropies of f (for n = 1 see [38]).

Exercice 2.24.

Prove (or disprove) that every n-expansive Axiom A dieomor-

phism of a closed manifold is expansive.

Exercice 2.25.

Are there n-expansive homeomorphisms of S 2 ? (the answer is


negative for n = 1, see Hiraide [42] and Lewowicz [59]).

Chapter 3

Positively expansive
measures
3.1

Introduction

Ergodic measures with positive entropy for continuous maps on compact metric spaces have been studied in the recent literature. For
instance, [14] proved that the set of points belonging to a proper
asymptotic pair (i.e. points whose stable classes are not singleton)
constitute a full measure set. Moreover, [43] proved that if f is a
homeomorphism with positive entropy h (f ) with respect to one of
such measures , then there is a full measure set A such that for all
x A there is a closed subset A(x) in the stable class of x satisfying
h(f 1 , A(x)) h (f ), where h(, ) is the Bowens entropy operation
[10]. We can also mention [27] which proved that every ergodic endomorphism on a Lebesgue probability space having positive entropy
on nite measurable partitions formed by continuity sets is pairwise
sensitive (see also Exercice 3.48).
In this chapter we introduce the notion of positively expansive
measure and prove that every ergodic measure with positive entropy
on a compact metric space is positively expansive. Using this result
we will prove that, on compact metric spaces, every stable class has
measure zero with respect to any ergodic measure with positive en40

[SEC. 3.2: DEFINITION

41

tropy (this seems to be new as far as we know). We also prove through


the use of positively expansive measures that every continuous map
on a compact metric space exhibiting countably many stable classes
has zero topological entropy (a similar result with dierent techniques
has been obtained in [45] but in the transitive case). Still in the compact case we prove that every continuous map which is Lyapunov
stable on its recurrent set has zero topological entropy too (this is
known but for one-dimensional maps [35], [81], [92]). Finally we use
expansive measures to give necessary conditions for a continuous map
on a complete separable metric space to be chaotic in the sense of Li
and Yorke [60]. Most results in this chapter were obtained in [6] and
[7].

3.2

Denition

In this chapter we introduce the notion of positively expansive measure. First we recall the following denition.
Denition 3.1. A continuous map f : X X of a metric space
X is positively expansive (c.f. [33]) if there is > 0 such that
for every pair of distinct points x, y X there is n N such that
d(f n (x), f n (y)) > . Equivalently, f is positively expansive if there
is > 0 such that (x) = {x}, where
(x) = {y X : d(f i (x), f i (y)) , i N}
(again we write f (x) to indicate dependence on f ).
This motivates the following denition
Denition 3.2. A positively expansive measure of a measurable map
f : X X is a Borel probability measure for which there is > 0
such that ( (x)) = 0 for all x X. The constant will be referred
to an positive expansivity constant of .
As in the invertible case we have that a measure is a positively expansive measure of f if and only if there is > 0 such that
( (x)) = 0 for -a.e. x X. An atomic measures cannot be
an expansive measure of any map and every non-atomic Borel probability measure is a positively expansive measure of any is positively
expansive map.

42

[CAP. 3: POSITIVELY EXPANSIVE MEASURES

Example 3.3. There are nonexpansive continuous maps on certain


compact metric spaces for which every non-atomic measure is expansive (e.g. an n-expansive homeomorphism with n 2). The homeomorphism f (x) = 2x in R exhibits positively expansive measures (e.g.
the Lebesgue measure) but not positively expansive invariant ones.
Contrary to what happen in the expansive case ([79], [77]), there
are innite compact metric spaces supporting homeomorphisms with
positively expansive measures (extreme cases will be discussed in Exercice 3.43). On the other hand, a necessary and sucient for a
measure to be positively expansive is given as in the homeomorphism
case.
We shall need a previous result stated as follows. Let f : X X
be a measurable map of a metric space X. Given x X, n N+ and
> 0 we dene
B[x, n, ] =

n1


f i (B[f i (x), ]).

(3.1)

i=0

A basic property of these sets is given below.


(x) =

B[x, n, ].

(3.2)

n=1

Since, in addition, B[x, m, ] B[x, n, ] for n m, we obtain


( (x)) = lim (B[x, , n])
n

for every x X and every Borel probability measure of X.


From this we obtain the pointwise convergence
= lim ,n
n

(3.3)

where , ,n : X R+ are the functions dened by


(x) = ( (x))

and

,n (x) = (B[x, , n]).

(3.4)

Moreover, is positively expansive if and only if there is > 0 such


that
for all x X.
(3.5)
lim inf (B[x, n, ]) = 0,
n

43

[SEC. 3.2: DEFINITION

It follows that if n N+ , then is a positively expansive measure


of f , if and only if it is a positively expansive measure of f n . The
proof of these assertions is analogous to the corresponding results for
homeomorphisms (see exercices 3.34 and 3.35).
Next we present the following lemma dealing with the measurability of the map .
Lemma 3.4. If f : X X is a continuous map of a compact metric
space X and is a nite Borel measure of X, then is a measurable
map for every > 0.
Proof. Fix > 0, n N+ and dene
Dn = {(x, y) X X : d(f i (x), f i (y)) ,

0 i n 1}.

Denote by B(Y ) the Borel -algebra associated to a topological space


X. Since f is continuous we have that Dn is closed in X X with respect to the product topology. From this we obtain Dn B(X X).
But since X is compact the product -algebra B(X) B(X) satises
B(X)B(X) = B(X X) (e.g. Lemma 6.4.2 in [17]). Therefore Dn
B(X)B(X). This allows us to apply the Fubini Theorem
(e.g. The
orem 3.4.1 in [17]) to conclude that the map x 
X

Dn (x, y)d(y)

is measurable, where Dn denotes the characteristic function of Dn .


But it follows from the denition of Dn that

Dn (x, y)d(y)
,n (x) =
X

so ,n is measurable, n N+ . It follows from (3.3) that is the


pointwise limit of measurable functions and so measurable.
As in the expansive case we have the following observation for
bijective maps f : X X, namely,
f ( (x)) (f (x)),

(x, ) X R+ .

Using it we obtain the elementary lemma below.


Lemma 3.5. Let f : X X be a homeomorphism of a metric space
X. If is an expansive measure with expansivity constant of f ,
then so is f1 .

44

[CAP. 3: POSITIVELY EXPANSIVE MEASURES

Following the proof of Lemma 1.17 and using (3.5) (instead of


Lemma 1.16) we can obtain the following result.
Lemma 3.6. If f : X X is a homeomorphism of a metric space
X, then every invariant measure of f which is the limit (with respect
to the weak-* topology) of a sequence of positively expansive probability measures with a common expansivity constant of f is positively
expansive for f .
Using this lemma we obtain the following result closely related to
Example 3.3.
Theorem 3.7. A homeomorphism of a compact metric space has
positively expansive probability measures if and only if it has positively
expansive invariant probability measures.
Proof. Let be a positively expansive measure with positive expansivity constant of a homeomorphism f : X X of a compact metric
space X. By Lemma 3.5 we have that f1 is a positively expansive
measure with positive expansivity constant of f . Therefore, fi is
a positively expansive measure with positively expansivity constant
of f (i N), and so,
n =

n1
1 i
f ,
n i=0

n N+

is a sequence of positively expansive probability measures of f with


common expansivity constant . As X is compact there is a subsequence nk such that nk converges to a Borel probability
measure . Since is clearly invariant for f 1 and f is a homeomorphism we have that is also an invariant measure of f . Then, we
can apply Lemma 3.6 to this sequence to obtain that is a positively
expansive measure of f .
An equivalent condition for positively -expansiveness is given
using the idea of positive generators as in Lemma 3.3 of [26]. Call
a nite open covering A of X positive -generator of f if for every
sequence {An : n N} A one has



n
f (Cl(An )) = 0.

nN

[SEC. 3.2: DEFINITION

45

As in the homeomorphism case we obtain the following proposition.


Proposition 3.8. Let f : X X be a continuous map of a compact
metric space X. Then, a Borel probability measure of X is a positively
invariant measure of f if and only if if f has a positive -generator.
We shall use this proposition to obtain examples of positively
expansive measures. If M is a closed manifold we call a dierentiable
map f : M M volume expanding if there are constants K > 0
and > 1 such that |det(Df n (x))| Kn for all x M and n
N. Denoting by Leb the Lebesgue measure we obtain the following
proposition.
Proposition 3.9. The Lebesgue measure Leb is a positively expansive measure of every volume expanding map of a closed manifold.
Proof. If f is volume expanding there are n0 N and > 1 such
that g = f n0 satises |det(Dg(x))| for all x M . Then, for all
x M there is x > 0 such that
Leb(g 1 (B[x, ])) 1 Leb(B[x, ]),

x M, 0 < < x .
(3.6)
Let be half of the Lebesgue number of the open covering {B(x, x ) :
x M } of M . By (3.6) any nite open covering of M by -balls
is a positive Leb-generator, so, Leb is positively expansive for g by
Proposition 3.8. Since g = f n0 we conclude that Leb is a positively
expansive measure of f (see the remark after (3.5)).
As in the homeomorphism case we obtain an equivalent condition
for positively expansiveness using the diagonal. Given a map g of a
metric space Y and a Borel probability in Y we say that I Y is
a -repelling set if there is a neighborhood U of I satisfying
({z Y : g n (z) U, n N}) = 0.
As in the homeomorphism case we can prove the following.
Proposition 3.10. Let f : X X be a continuous map of a compact
metric space X. Then, a Borel probability measure of X is a positively
expansive for f if and only if the diagonal is a 2 -repelling set of
f f.

46

[CAP. 3: POSITIVELY EXPANSIVE MEASURES

We shall use the following useful characterization of positively


expansive measures which is analogous to the expansive case (c.f.
Lemma 1.20).
Lemma 3.11. A Borel probability measure is positively expansive
for a measurable map f if and only if there is > 0 such that
( (x)) = 0,

-a.e. x X.

(3.7)

This lemma together with the corresponding denition for expansive maps suggests the following.
Denition 3.12. A positively expansive constant of a Borel probability measure is a constant > 0 satisfying (3.7).

3.3

Properties

In this section we select the properties of positively expansive measures we shall use later one. For the rst one we need the following
denition.
Denition 3.13. Given a map f : X X and p X we dene
W s (p), the stable set of p, as the set of points x for which the pair
(p, x) is asymptotic, i.e.,


W s (p) = x X : lim d(f n (x), f n (p)) = 0 .
n

By a stable class we mean a subset equals to W s (p) for some p X.


The following shows that every stable class is negligible with respect to any expansive invariant measure.
Proposition 3.14. The stable classes of a measurable map have measure zero with respect to any positively expansive invariant measure.
Proof. Let f : X X a measurable map and be a positively expansive invariant measure. Denoting by B[, ] the closed ball operation
one gets


  
1
s
k
k
f
B f (p),
.
W (p) =
i
+
iN

jN kj

47

[SEC. 3.3: PROPERTIES

As clearly
 


f k B f k (p),

jN kj

1
i+1

 
jN kj



1
f k B f k (p),
,
i

(i N+ ) we obtain
(W s (p)) lim


jN


kj



1
.
f k B f k (p),
i

(3.8)

On the other hand,



kj





1
f k B f k (p),
= f j 1i (f j (p))
i

so


kj



1
=
f k B f k (p),
i





f j 1i (f j (p)) = 1i (f j (p))


since is invariant. Then, taking i large, namely, i > 1 where


 is a expansivity
constant of (c.f. Denition 3.12) we obtain

1i (f j (p)) = 0 so


kj



1
f k B f k (p),
= 0.
i

Replacing in (3.8) we get the result.


For the second property we will use the following denition [35].
Denition 3.15. A map f : X X is said to be Lyapunov stable on
A X if for any x A and any  > 0 there is a neighborhood U (x)
of x such that d(f n (x), f n (y)) <  whenever n 0 and y U (x) A.

48

[CAP. 3: POSITIVELY EXPANSIVE MEASURES

(Notice the dierence between this denition and the corresponding one in [81].) The following implies that measurable sets where the
map is Lyapunov stable are negligible with respect to any expansive
measure (invariant or not).
Proposition 3.16. If a measurable map of a separable metric space
is Lyapunov stable on a measurable set A, then A has measure zero
with respect to any positively expansive measure.
Proof. Fix a measurable map f : X X of a separable metric space
X, a positively expansive measure and > 0. Since is regular
there is a closed subset C A such that
(A \ C) .
Let us compute (C).
Fix a positive expansivity constant  of (c.f. Denition 3.12).
Since f is Lyapunov stable on A and C A for every x C there is
a neighborhood U (x) such that
d(f n (x), f n (y)) < 

n N, y U (x) C.

(3.9)

On the other hand, C is separable (since X is) and so Lindel


of with
the induced topology. Consequently, the open covering {U (x) C :
x C} of C admits a countable subcovering {U (xi ) C : i N}.
Then,

(U (xi ) C) .
(3.10)
(C)
iN

Now x i N. Applying (3.9) to x = xi we obtain U (xi ) C


 (xi ) and then (U (xi ) C) ( (x)) = 0 since  is a positive
expansivity constant. As i is arbitrary we obtain (C) = 0 by (3.10).
To nish we observe that
(A) = (A \ C) + (C) = (A \ C)
and so (A) = 0 since is arbitrary. This ends the proof.
From these propositions we obtain the following corollary. Recall
that the recurrent set of f : X X is dened by R(f ) = {x X :
x f (x)}, where


nk
f (x) = y X : y = lim f (x) for some sequence nk .
k

49

[SEC. 3.4: APPLICATIONS

Corollary 3.17. A measurable map of a separable metric space which


either has countably many stable classes or is Lyapunov stable on its
recurrent set has no positively expansive invariant measures.
Proof. First consider the case when there are countably many stable
classes. Suppose by contradiction that there exists a positively expansive invariant measure. Since the collection of stable classes is a
partition of the space it would follow from Proposition 3.14 that the
space has measure zero which is absurd.
Now consider the case when the map f is Lyapunov stable on
R(f ). Again suppose by contradiction that there is a positively expansive invariant measure . Since is invariant we have supp()
R(f ) by Poincare recurrence. However, since f is Lyapunov stable on
R(f ) we obtain (R(f )) = 0 from Proposition 3.16 so (supp())) =
(R(f )) = 0 which is absurd. This proves the result.

3.4

Applications

We start this section by proving that positive entropy implies expansiveness among ergodic invariant measures for continuous maps on
compact metric spaces. Afterward we include some short applications.
To star with we introduce the following basic result due to Brin
and Katok [18]. Let be an invariant measure of a measurable map
f : X X of a metric space X. The entropy of with respect to f
is dened by
h (f ) = sup{h (f, P ) : P is a nite measurable partition of X},
where

1
() log ()
n n

h (f, P ) = lim

Pn1

and Pn is the pullback partition of P under f n .


Theorem 3.18 (Brin-Katok Theorem). If is a non-atomic ergodic invariant measure of a continuous map f : X X of a compact
metric space, then
sup lim inf
>0 n

log((B[x, n, ]))
= h (f ),
n

-a.e. x X.

50

[CAP. 3: POSITIVELY EXPANSIVE MEASURES

Next we state the main result of this section.


Theorem 3.19. Every ergodic invariant probability measure with
positive entropy of a continuous map on a compact metric space is
positively expansive.
Proof. Let be an ergodic invariant measure with positive entropy
h (f ) > 0 of a continuous map f : X X on a compact metric
space X. Fix > 0 and dene
X = {x X : ( (x)) = 0}.
Clearly X = 1
(0) (where is dened in (3.4) and so X is measurable by Lemma 3.4. Then, we are left to prove by Lemma 3.11
that there is > 0 such that (X ) = 1.
Fix x X. It follows from the denition of (x) that (x)
f 1 ( (f (x))) so
( (x)) ( (f (x)))
since is invariant. Then, ( (x)) = 0 whenever x f 1 (X )
yielding
f 1 (X ) X .
Denote by AB the symmetric dierence of the sets A, B. Since
(f 1 (X )) = (X ) the above implies that X is essentially invariant, i.e., (f 1 (X )X ) = 0. Since is ergodic we conclude that
(X ) {0, 1} for all > 0. Then, we are left to prove that there is
> 0 such that (X ) > 0. To nd it we proceed as follows.
For all > 0 we dene the map : X IR {},
(x) = lim inf
n

Take h =

h (f )
2

Xm

log (B[x, n, ])
.
n

(thus h > 0) and dene




= x X : m1 (x) > h ,

m IN + .

Notice that (x)  (x) whenever 0 < <  .



From this it follows that X m X m for m m and further



x X : sup (x) = h (f )
X m.
>0

mIN +

51

[SEC. 3.4: APPLICATIONS

Then,



x X : sup (x) = h (f )
>0

lim (X m ).
m

But is non-atomic for it is ergodic invariant with positive entropy.


So, the Brin-Katok Theorem implies



x X : sup (x) = h (f )
=1
>0

yielding
lim (X m ) = 1.

Consequently, we can x m IN + such that


(X m ) > 0.
1
works.
We shall prove that = m
m
Let us take x X . It follows from the denition of X m that
(B[x, n, ]) < ehn for all n large. Since h > 0 we conclude that
limn (B[x, n, ]) = 0. Since ,n (x) = (B[x, n, ]) we conclude
from (3.3) that ( (x)) = 0 thus x X . As x X m is arbitrary
we obtain X m X whence

0 < (X m ) (X )
and the proof follows.
The converse of the above theorem is false, i.e., a positively expansive measure may have zero entropy even in the ergodic invariant
case. A counterexample is as follows.
Example 3.20. There are continuous maps in the circle exhibiting
ergodic invariant measures with zero entropy which, however, are positively expansive.
Proof. Since all circle homeomorphisms have zero topological entropy
it remains to prove that every Denjoy map h exhibits positively expansive measures. As is well-known h is uniquely ergodic and the
support of its unique invariant measure is a minimal set, i.e., a

52

[CAP. 3: POSITIVELY EXPANSIVE MEASURES

set which is minimal with respect to the property of being compact


invariant. We shall prove that this measure is positively expansive.
Denote by E the support of . It is well known that E is a Cantor
set. Let be half of the length of the biggest interval I in the complement S 1 E of E and take 0 < < /2. Fix x S 1 and denote
by Int() the interior operation. We claim that Int( (x)) E = .
Otherwise, there is some z Int( (x)) E. Pick w I (thus
w E). Since E is minimal there is a sequence nk such that
hnk (w) z. Since is a nite measure, the interval sequence
{hn (I) : n IN } is disjoint, we have that the length of the intervals
hnk (I) 0 as k . It turns out that there is some integer k
such that hnk (I) (x).
From this and the fact that h( (x)) (h(x)) one sees that
I B[hnk (x), ] which is clearly absurd because the length of I is
greather than > 2. This contradiction proves the claim. Since
(x) is either a closed interval or {x} the claim implies that (x)
E = (x) E consists of at most two points. Since is clearly
non-atomic we conclude that ( (x)) = 0. Since x S 1 is arbitrary
we are done.
A rst application of Theorem 3.19 is as follows.
Theorem 3.21. The stable classes of a continuous map of a compact
metric space have measure zero with respect to any ergodic invariant
measure with positive entropy.
Proof. In fact, since these measures are positively expansive by Theorem 3.19 we obtain the result from Proposition 3.14.
We can also use Theorem 3.19 to compute the topological entropy
of certain continuous maps (for the related concepts see [3] or [89]).
As a motivation let us mention the known facts that both transitive
continuous maps with countably many stable classes on compact metric spaces and continuous maps of the interval or the circle which are
Lyapunov stable on their recurrent sets have zero topological entropy
(see Corollary 2.3 p. 263 in [45], [35], Theorem B in [81] and [92]).
Indeed we improve these result in the following way.

53

[SEC. 3.4: APPLICATIONS

Theorem 3.22. A continuous map of a compact metric space which


either has countably many stable classes or is Lyapunov stable on its
recurrent set has zero topological entropy.
Proof. If the topological entropy were not zero the variational principle [89] would imply the existence of ergodic invariant measures with
positive entropy. But by Theorem 3.19 these measures are positively
expansive against Corollary 3.17.
Example 3.23. An example satisfying the rst part of Theorem 3.22
is the classical pole North-South dieomorphism on spheres. In fact,
the only stable sets of this dieomorphism are the stable sets of the
poles. The Morse-Smale dieomorphisms [40] are basic examples
where these hypotheses are fullled.
Now we use positively expansive measures to study the chaoticity
in the sense of Li and Yorke [60]. Recall that if 0 a -scrambled
set of f : X X is a subset S X satisfying
lim inf d(f n (x), f n (y)) = 0
n

and

lim sup d(f n (x), f n (y)) >


n

(3.11)
for all dierent points x, y S. The following result relates scrambled
sets with positively expansive measures.
Theorem 3.24. A continuous map of a Polish metric space carrying
an uncountable -scrambled set for some > 0 also carries positively
expansive probability measures.
Proof. Let X a Polish metric space and f : X X be a continuous
map carrying an uncountable -scrambled set for some > 0. Then,
by Theorem 16 in [13], there is a closed uncountable -scrambled set
S. As S is closed and X is Polish we have that S is also a Polish metric
space with respect to the induced metric. As S is uncountable we have
from [73] that there is a non-atomic Borel probability measure in S.
Let be the Borel probability induced by in X, i.e., (A) = (AS)
for all Borelian A X. We shall prove that this measure is positively
expansive. If x S and y (x) S we have that x, y S and
2
d(f n (x), f n (y)) 2 for all n N therefore x = y by the second
inequality in (3.11). We conclude that (x) S = {x} for all x S.
2

54

[CAP. 3: POSITIVELY EXPANSIVE MEASURES

As is non-atomic we obtain ( (x)) = ( (x)S) = ({x}) = 0


2
2
for all x S. On other hand, it is clear that every open set which
does not intersect S has -measure 0 so is supported in the closure
of S. As S is closed we obtain that is supported on S. We conclude
that ( (x)) = 0 for -a.e. x X, so, is positively expansive by
2
Lemma 3.11.
Corollary 3.25. Every homeomorphism of a compact metric space
carrying an uncountable -scrambled set for some also carries positively expansive invariant probability measures.
Proof. Every compact metric space is Polish so Theorem 3.24 yields
positively expansive probability measures. Now apply Theorem 3.7.
Now recall that a continuous map is Li-Yorke chaotic if it has an
uncountable 0-scrambled set.
Until the end of this section M will denote either the interval
I = [0, 1] or the unit circle S 1 .
Corollary 3.26. Every Li-Yorke chaotic map in M carries positively
expansive measures.
Proof. Theorem in p. 260 of [31] together with theorems A and B in
[57] imply that every Li-Yorke chaotic map in M has an uncountable
-scrambled set for some > 0. Then, we obtain the result from
Theorem 3.24.
It follows from Example 3.20 that there are continuous maps with
zero topological entropy in the circle exhibiting positively expansive
invariant measures. This leads to the question whether the same
result is true on compact intervals. The following consequence of the
above corollary gives a partial positive answer for this question.
Example 3.27. There are continuous maps with zero topological entropy in the interval carrying positively expansive measures.
Indeed, by [47] there is a continuous map of the interval, with zero
topological entropy, exhibiting a -scrambled set of positive Lebesgue
measures for some > 0. Since sets with positive Lebesgue measure

55

[SEC. 3.4: APPLICATIONS

are uncountable we obtain a positively expansive measure from Theorem 3.24.


Another interesting example is the one below.
Example 3.28. The Lebesgue measure is an ergodic invariant measure with positive entropy of the tent map f (x) = 1 |2x 1| in I.
Therefore, this measure is positively expansive by Theorem 3.19.
It follows from this example that there are continuous maps in
I carrying positively expansive measures with full support (i.e.
supp() = I). These maps also exist in S 1 (e.g. an expanding map).
Now, we prove that Li-Yorke and positive topological entropy are
equivalent properties among these maps in I. But previously we
need a result based on the following well-known denition.
A wandering interval of a map f : M M is an interval J M
such that f n (J) f m (J) = for all dierent integers n, m N and
no point in J belongs to the stable set of some periodic point.
Lemma 3.29. If f : M M is continuous, then every wandering
interval has measure zero with respect to every positively expansive
measure.
Proof. Let J a wandering interval and be a positively expansive
measure with expansivity constant  (c.f. Denition 3.12). To prove
(J) = 0 it suces to prove Int(J) supp() = 0 since is nonatomic. As J is a wandering interval one has limn |f n (J)| = 0,
where | | denotes the length operation.
From this there is a positive integer n0 satisfying
|f n (J)| < ,

n n0 .

(3.12)

Now, take x Int(J). Since f is clearly uniformly continuous and


n0 is xed we can select > 0 such that B[x, ] Int(J) and
|f n (B[x, ])| <  for 0 n n0 . This together with (3.12) implies |f n (x) f n (y)| <  for all n N therefore B[x, ]  (x) so
(B[x, ]) = 0 since  is an expansivity constant. Thus x
supp()
and we are done.
From this we obtain the following corollary.

56

[CAP. 3: POSITIVELY EXPANSIVE MEASURES

Corollary 3.30. A continuous map carrying positively expansive


measures with full support of the circle or the interval has no wandering intervals. Consequently, a continuous map of the interval carrying positively expansive measures with full support is Li-Yorke chaotic
if and only if it has positive topological entropy.
Proof. The rst part is a direct consequence Lemma 3.29 while, the
second, follows from the rst since a continuous interval map without
wandering intervals is Li-Yorke chaotic if and only if it has positive
topological entropy [82].

3.5

The smooth case

Now we turn our attention to smooth ergodic theory. The motivation


is the well-known fact that a dieomorphism restricted to a hyperbolic basic set is expansive. In fact, it is tempting to say that every
hyperbolic ergodic measures of a dieomorphism is positively expansive (or at least expansive) but the Dirac measure supported on a
hyperbolic periodic point is a counterexample. This shows that some
extra hypotheses are necessary for a hyperbolic ergodic measure to
be positively expansive. Indeed, by the results above, we only need
to recognize which conditions imply positive entropy. Let us state
some basic denitions in order to present our result.
Assume that X is a compact manifold and that f is a C 1 dieomorphism. We say that point x X is a regular point whenever there
are positive integers s(x) and numbers {1 (x), , s(x) (x)} IR
(called Lyapunov exponents) such that for every v Tx M \ {0} there
is 1 i s(x) such that
1
log Df n (x)v = i (x).
n
An invariant measure is called hyperbolic if there is a measurable
subset A with (A) = 1 such that i (x)
= 0 for all x A and all
1 i s(x).
On the other hand, the Eckmann-Ruelle conjecture [9] asserts that
every hyperbolic ergodic measure is exac-dimensional, i.e., the limit
below
(B(x, r))
d(x) = lim+
r
r0
lim

[SEC. 3.6: EXERCICES

57

exists and is constant -a.e. x X. This constant is the so-called


dimension of .
With these denitions we can state the following result.
Theorem 3.31. Let f be a C 2 dieomorphism of a compact manifold.
1. Every hyperbolic ergodic measure of f which either has positive
dimension or is absolutely continuous with respect to Lebesgue
is positively expansive.
2. If f has a non-atomic hyperbolic ergodic measure, then f also
has a positively expansive ergodic invariant measure.
Proof. Let us prove (1). First assume that the measure has positive
dimension. As noticed in [9] p. 761 Theorem C p. 544 in [58] implies
that if the entropy vanishes, then the stable and unstable dimension
of the measure also do. In such a case we have from Theorem F
p. 548 in [58] that the measure has zero dimension, a contradiction.
Therefore, the measure has positive entropy and then we are done by
Theorem 3.19.
Now assume that the measure is absolutely continuous with respect to the Lebesgue measure. Then, it is non-atomic so the argument in the proof of Theorem 4.2 p. 167 in [56] implies that it
has at least one positive Lyapunov exponent. Therefore, the Pesin
formula (c.f. p. 139 in [52]) implies positive entropy so we are done
by Theorem 3.19.
To prove (2) we only have to see that Corollary 4.2 in [52] implies
that every dieomorphism as in the statement of (2) has positive
topological entropy. Then, we are done by the variational principle
and Theorem 3.19 (see Exercice 3.39).

3.6

Exercices

Exercice 3.32.

Prove that the Lebesgue measure of S 2 is an expansive measure


of the Bernoulli dieomorphism in S 2 found in [53] (therefore Corollary 1.41 is false

for S 2 instead of S 1 ). Is such a dieomorphism measure-expansive?

58

[CAP. 3: POSITIVELY EXPANSIVE MEASURES

Exercice 3.33.

Is it true that every continuous map f : X X exhibiting

positively expansive probability measures of a compact metric space also exhibits


positively expansive invariant measures?

Exercice 3.34.

Let f : X X be a measurable map of a metric space X.


Prove that a Borel probability measure of X is positively expansive for f if and
only if if there is > 0 such that
for all x X,

lim inf (B[x, n, ]) = 0,


n

where B[x, n, ] is dened in 3.1.

Exercice 3.35. Prove the equivalence of the following properties for every continuous map f : X X of compact metric space and every Borel probabiltity
measure of X:
is positively expansive for f ;
there is n N+ such that is positively expansive for f n ;
is positively expansive for f n , n N+ .

Exercice 3.36.

Prove that the constant map cannot have positively expansive

measures.

Exercice 3.37.

Prove lemmas 3.5, 3.11, 3.6 and Proposition 3.10.

Exercice 3.38.

Prove that e Borel probability measure is positively expansive


for a measurable map f : X X of a metric space X if and only if there are > 0
and a negligible set X0 of X such that ( (x)) = 0 for every x X0 (negligible
means that (A) = 0 for every measurable subset A X0 ).

Exercice 3.39.

Prove that every continuous map of a compact metric space


f : X X satises the variational principle,
h(f ) =

sup

M
exp (f )

h (f ),

where Mexp (f ) denotes the space of expansive invariant probability measures of f


(of course, with the supremum being zero if Mexp (f ) = ).

Exercice 3.40.

Following [21] we say that a Borel measure of a metric space

X is almost expansive for a Borel isomorphism f : X X if there is > 0 such


that (x) = {x} for -a.e. x X. Find examples of homeomorphisms of compact
metric spaces exhibiting expansive ergodic invariant measures which are not almost
expansive.

Exercice 3.41.

Prove that a circle homeomorphism exhibits positively expansive

measures if and only if it is Denjoy.

59

[SEC. 3.6: EXERCICES

Exercice 3.42.

Investigate the parameter values 0 1 for which the

Lebesgue measure is positively expansive for the map g (x) = (1 |2x 1|) of the
unit interval I. Analogously for the family f (x) = x(1 x) , 0 4.

Exercice 3.43.

Call a continuous map f : X X of a non-atomic metric


space X positively measure-expansive if every non-atomic Borel measure is positively
expansive for f . Find examples of positively measure-expansive homeomorphisms of
non-atomic compact metric spaces.

Exercice 3.44.

Find a homeomorphism of a compact non-atomic metric space


which is positively measure-expansive (and so measure-expansive) but not expansive.

Exercice 3.45.

Prove that there are no Li-Yorke chaotic homeomorphisms of


the circle. Conclude that there are continuous maps of compact metric spaces with
positively expansive measures which are not Li-Yorke chaotic.

Exercice 3.46. Does every Li-Yorke chaotic map of a compact metric space
carry positively expansive measures?
Exercice 3.47.

Are there dieomorphisms of closed manifolds exhibiting nonatomic hyperbolic measure which are neither expansive nor positively expansive?

Exercice 3.48. A measurable map f : X X of a metric is called pairwise


sensitive for a Borel measure if there is > 0 such that
2 ({(x, y) X X : n N such that d(f n (x), f n (y)) }) = 1
(c.f. [27]). Prove that a Borel probability measure of X is positively expansive for
f if and only if f is pairwise sensitive for .

Chapter 4

Measure-sensitive maps
4.1

Introduction

In this chapter we will try to extend the notion of measure expansivity from metric to measurable spaces. For this we introduce
the auxiliary denition of measure-sensitive partitions and measuresensitive spaces. We prove that every non-atomic standard probability spaces is measure-sensitive and that every measure-sensitive
probability spaces is non-atomic. With this concept we introduce the
notion of measure-sensitive partition which will play a role similar
to the expansivity constant for expansive maps. We prove that in
a non-atomic probability space every strong generator is a measuresensitive partition but not conversely (results about strong generators
can be found in [41], [48], [69], [70] and [71]). We exhibit examples of measurable maps in non-atomic probability spaces carrying
measure-sensitive partitions which are not strong generators. Motivated by these examples we shall study the measure-sensitive maps(1 )
i.e. measurable maps on measure spaces carrying measure-sensitive
partitions. Indeed, we prove that every measure-sensitive map is aperiodic and also, in the probabilistic case, that its corresponding space
is non-atomic.
From this we obtain a characterization of nonsingular countable
1 Called

measure-expansive maps in [64]

60

[SEC. 4.2: MEASURE-SENSITIVE SPACES

61

to one measure-sensitive mappings on non-atomic Lebesgue probability spaces as those having strong generators. Furthermore, we
prove that every ergodic measure-preserving map with positive entropy is a probability space is measure-sensitive (thus extending a
result in [27]). As an application we obtain some properties for ergodic measure-preserving maps with positive entropy (c.f. corollaries
4.14 and 4.20). A reference for the results in this chapter is [64].

4.2

Measure-sensitive spaces

Hereafter the term countable will mean either nite or countably innite.
A measure space is a triple (X, B, ) where X is a set, B is a algebra of subsets of X and is a positive measure in B. A probability
space is one for which (X) = 1.
A partition is a disjoint collection P of nonempty measurable sets
whose union is X. We allow () = 0 for some P . Given
partitions P and Q we write P Q to mean that each member of Q
is contained in some member of P (mod 0). A sequence of partitions
{Pn : n N} (or simply Pn ) is increasing if Pi Pj for i j.
Motivated by the concept of Lebesgue sequence of partitions (c.f.
p. 81 in [61]) we introduce the following denition.
Denition 4.1. A measure-sensitive sequence of partitions of a measure space (X, B, ) is an increasing sequence of countable partitions
Pn such that



n = 0

nN

for all sequence of measurable sets n satisfying n Pn , n N. A


measure-sensitive space is a measure space carrying measure-sensitive
sequences of partitions.
Let us present a sucient condition for sequences of partitions to
be measure-sensitive. Recall that the join of nitely many partitions
P0 , , Pn is the partition dened by
 n

n


Pk =
k : k Pk , 0 k n .
k=0

k=0

62

[CAP. 4: MEASURE-SENSITIVE MAPS

Certainly
Pn =

n


f k (P ),

n N,

(4.1)

k=0

denes an increasing sequence of countable partitions satisfying


Pn (x) =

n


f k (P (f k (x)),

x X.

k=0

Since for all x X one has


{y X : f n (y) P (f n (x)),

f n (P (f n (x))) =

n=0

n N} =
Pn (x),

n=0

we obtain that the identity below


lim sup () = 0

n Pn

(4.2)

is sucient condition for an increasing sequence Pn of countable partitions to be measure-sensitive. It is also necessary in probability
spaces (see Exercice 4.27).
Let us state basic properties of the measure-sensitive spaces. For
this recall that a measure space is non-atomic if it has no atoms, i.e.,
measurable sets A of positive measure satisfying (B) {0, (A)}
for every measurable set B A. Recall that a standard probability
space is a probability space (X, B, ) whose underlying measurable
space (X, B) is isomorphic to a Polish space equipped with its Borel
-algebra (e.g. [1]).
The class of measure-sensitive spaces is broad enough to include
all non-atomic standard probability spaces. Precisely we have the
following proposition.
Proposition 4.2. Every non-atomic standard probability spaces is
measure-sensitive.
Proof. It is well-known that if (X, B, ) is a non-atomic standard
probability space, then there are a measurable subset X0 X with

63

[SEC. 4.2: MEASURE-SENSITIVE SPACES

(X \ X0 )= 0 and a sequence of countable partitions Qn of X0


such that nN n contains at most one point for every sequence of
measurable sets n in X0 satisfying n Qn , n N (c.f. [61] p.
81). Dening Pn = {X \ X0 } Qn we obtain an increasing sequence
of countable partitions of (X, B, ). It suces to prove that this
sequence is measure-sensitive. For this take a xed (but arbitrary)
sequence of measurable sets n of X with n Pn for all n N. It
follows from the denition of Pn that either n = X \ X0 for
 some
n N, or, n Qn for all n N. Then, the intersection nN n
either is contained in X \ X0 or reduces to a single measurable point.
Since both X \ X0 and the measurable points have

zero (for
 measure
non-atomic spaces are diuse [10]) we obtain nN n = 0. As n
is arbitrary we are done.
Although measure-sensitive probability spaces need not be standard (Exercice 4.26) we have that all of them are non-atomic. Indeed,
we have the following result of later usage.
Proposition 4.3. Every measure-sensitive probability spaces is nonatomic.
Proof. Suppose by contradiction that a measure-sensitive probability
space (X, B, ) has an atom A. Take a measure-sensitive sequence of
partitions Pn . Since A is an atom one has that n N !n Pn
such that (A n ) > 0 (and so (A n ) = (A)). Notice that
(n n+1 ) > 0 for, otherwise, (A) (A (n n+1 )) = (A
n ) + (A n+1 ) = 2(A) which is impossible in probability spaces.
Now observe that n Pn and Pn Pn+1 , so, there is L Pn+1
such that


= 0.
(4.3)
n 
If n+1
yielding


L

= we would have n n+1 = n n+1 \

(n n+1 ) = n n+1 \


L

n \


L


L

= 0

64

[CAP. 4: MEASURE-SENSITIVE MAPS

which is absurd. Hence n+1



and then n+1 L for


=
L

Pn+1 is a partition and n+1 Pn+1 . Using (4.3) we obtain n+1 n


(mod 0) so A n+1 A n (mod 0) for all n N+ .
From this and well-known properties of probability spaces we obtain






n =
(A n ) = lim (A n ) = (A) > 0.
A
nN

nN


But Pn is measure-sensitive
and n Pn , n N, so nN n = 0



yielding A nN n = 0 which contradicts the above expression.
This contradiction yields the proof.

4.3

Measure-sensitive maps

Let (X, B) be a measure space. If f : X X is measurable and


k N we dene for every partition P the pullback partition f k (P ) =
{f k () : P } which is countable if P is.
Denition 4.4. A measure-sensitive partition of a measurable map
f : X X is a countable partition P satisfying
({y X : f n (y) P (f n (x)),

n N}) = 0,

x X,

(4.4)

where P (x) stands for the element of P containing x X.


The basic examples of measure-sensitive partitions are given as
follows. A strong generator of a measurable map f : X X is a
countable partition P for which the smallest -algebra of B containing

k
(P ) equals B (mod 0) (see [69]).
kN f
The result below is the central motivation of this chapter.
Theorem 4.5. Every strong generator of a measurable map f in a
non-atomic probability space is a measure-sensitive partition of f .
Proof. Let P be a strong generator of a measurable map f : X X
in a non-atomic probability space (X, B, ). Then, the sequence (4.1)
generates B (mod 0).

65

[SEC. 4.3: MEASURE-SENSITIVE MAPS

From this and Lemma 5.2 p. 8 in [61] we obtain that the set of
all nite unions of elements of these partitions is everywhere dense in
the measure algebra associated to (X, B, ). Consequently, Lemma
9.3.3 p. 278 in [10] implies that the sequence (4.1) satises (4.2) and
then (4.4) holds.
We shall see in Example 4.13 that the converse of this theorem is
false, i.e., there are certain measurable maps in non-atomic probability spaces carrying measure-sensitive partitions which are not strong
generators. These examples motivates the study of measure-sensitive
partitions for measurable maps in measure spaces.
The following equivalence relates both measure-sensitive partitions for maps and measure-sensitive sequences of partitions of measurable spaces
Lemma 4.6. The following properties are equivalent for measurable
maps f : X X and countable partitions P on measure spaces
(X, B, ):
(i) The sequence Pn in (4.1) is measure-sensitive for X.
(ii) The partition P is measure-sensitive for f .
(iii) The partition P satises
({y X : f n (y) P (f n (x)),

n N}) = 0, -a.e. x X.

Proof. Previously we state some notation. Given a partition P and


f : X X measurable we dene
P (x) = {y X : f n (y) P (f n (x)), n N},
Notice that
P (x) =

Pn (x)

x X.

(4.5)

nN+

and
Pn (x) =

n

i=0

f i (P (f i (x)))

(4.6)

66

[CAP. 4: MEASURE-SENSITIVE MAPS

so each P (x) is a measurable set. For later use we keep the following
identity

 n

i
f (P ) (x) = Pn (x),
x X.
(4.7)
i=0

Clearly (4.4) (resp. (iii)) is equivalent to (P (x)) = 0 for every


x X (resp. for -a.e. x X).
First we prove that (i) implies (ii). Suppose that the sequence
(4.1) is measure-sensitive
and x x X. By (4.5) and (4.7) we have

Pn is
P (x) = nN n where n = Pn (x) Pn . As the sequence


measure-sensitive we obtain (P (x)) = nN n = 0 proving
(ii). Conversely, suppose that (ii) holds and let 
n be a sequence of
measurable sets with n Pn for all n. Take y nN n . It follows
P (x) by
that y Pn(x) for all n N whence y 

(4.1). We conclude that nN n P (x) therefore nN n (P (x)) = 0


proving (i).
To prove that (ii) and (iii) are equivalent we only have to prove
that (iii) implies (i). Assume by contradiction that P saties (iii)
but not (ii). Since is a probability and (3) holds the set X  =
{x X : (P (x)) = 0} has measure one. Since (ii) does not hold
there is x X such that (P (x)) > 0. Since is a probability
and X  has measure one we would have P (x) X 
= so there
is y P (x) such that (P (y)) = 0. But clearly the collection
{P (x) : x X} is a partition (for P is) so P (x) = P (y) whence
(P (x)) = (P (y)) = 0 which is a contradiction. This ends the
proof.
Recall that a measurable map f : X X is measure-preserving if
f 1 = . Moreover, it is ergodic if every measurable invariant set A
(i.e. A = f 1 (A) (mod 0)) satises either (A) = 0 or (X \ A) = 0;
and totally ergodic if f n is ergodic for all n N+ .
Example 4.7. If f is a totally ergodic measure-preserving map of a
probability space, then every countable partition P with 0 < () < 1
for some P is measure-sensitive with respect to f (this follows
from the equivalence (iii) in Lemma 4.6 and Lemma 1.1 p. 208 in
[61]).

67

[SEC. 4.3: MEASURE-SENSITIVE MAPS

Hereafter we x a measure space (X, B, ) and a measurable map


f : X X. We shall not assume that f is measure-preserving unless
otherwise stated.
Using the Kolmogorov-Sinais entropy we obtain sucient conditions for the measure-sensitivity of a given partition. Recall that the
entropy of a nite partition P is dened by

() log ().
H(P ) =
P

The entropy of a nite partition P with respect to a measurepreserving map f is dened by


h(f, P ) = lim

1
H(Pn1 ).
n

Then, we have the following lemma.


Lemma 4.8. A nite partition with nite positive entropy of an ergodic measure-preserving map f in a probability space is a measuresensitive partition of f .
Proof. Since f is ergodic, the Shannon-Breiman Theorem (c.f. [61]
p. 209) implies that the partition P (say) satises
1
log((Pn (x))) = h(f, P ),
n n

lim

-a.e. x X,

(4.8)

where Pn (x) is as in (4.6). On the other hand, Pn+1 (x) Pn (x) for
all n so (4.5) implies
(P (x)) = lim (Pn (x)),
n

x X.

(4.9)

But h(f, P ) > 0 so (4.8) implies that (Pn (x)) goes to zero for -a.e.
x X. This together with (4.9) implies that P satisfy the equivalence
(iii) in Lemma 4.6 so P is measure-sensitive.
It follows at once from Lemma 4.6 that measure-sensitive maps
only exist on measure-sensitive spaces. Consequently we obtain the
following result from Proposition 4.3.

68

[CAP. 4: MEASURE-SENSITIVE MAPS

Theorem 4.9. Every probability space carrying measure-sensitive


maps is non-atomic.
A simple but useful example is as follows.
Example 4.10. The irrational rotations in the circle are measuresensitive maps with respect to the Lebesgue measure. This follows
from Example 4.7 since all such maps are measure-preserving and
totally ergodic.
On the other hand, it is not dicult to nd examples of measuresensitive measure-preserving maps which are not ergodic. These examples together with Example 4.10 suggest the question whether an
ergodic measure-preserving map is measure-sensitive. However, the
answer is negative by the following example.
Example 4.11. If (X, B, ) is a measure space with B = {X, }, then
no map is measure-sensitive although they are all ergodic measurepreserving.
In spite of this we can give conditions for the measure-expansivity
of ergodic measure-preserving maps as follows.
Recall that the entropy (c.f. [61], [89]) of f is dened by
h(f ) = sup{h(f, Q) : Q is a nite partition of X}.
We obtain a result closely related to Theorem 3.19 and Theorem 3.1
in [27].
Theorem 4.12. Every ergodic measure-preserving maps with positive entropy of a probability space is measure-sensitive.
Proof. Let f be one of such a map with entropy h(f ) > 0. We can
assume that h(f ) < . It follows that there is a nite partition Q
n1
with 0 < h(f, Q) < . Taking P = i=0 f i (Q) with n large we
obtain a nite partition with nite positive entropy since h(f, P ) =
h(f, Q) > 0. It follows that P is measure-sensitive by Lemma 4.8
whence f is measure-sensitive by denition.
A rst consequence of the above result is that the converse of
Theorem 4.5 is false.

[SEC. 4.4: APERIODICITY

69

Example 4.13. Let f : X X be a homeomorphism with positive topological entropy of a compact metric space X. By the variational principle [89] there is a Borel probability measures with
respect to which f is an ergodic measure-preserving map with positive entropy. Then, by Theorem 4.12, f carries a measure-sensitive
partition which, by Corollary 4.18.1 in [89], cannot be a strong generator. Consequently, there are measurable maps in certain non-atomic
probability spaces carrying measure-sensitive partitions which are not
strong generators.
On the other hand, it is also false that ergodic measure-sensitive
measure-preserving maps on probability spaces have positive entropy.
The counterexamples are precisely the irrational circle rotations (c.f.
Example 4.10). Theorems 4.9 and 4.12 imply the probably wellknown result below.
Corollary 4.14. Every probability spaces carrying ergodic measurepreserving maps with positive entropy is non-atomic.

4.4

Aperiodicity

In this section we analyse the aperiodicity of measure-sensitive maps.


According to [69] a measurable map f is aperiodic whenever for all
n N+ if n N+ and f n (x) = x on a measurable set A, then
(A) = 0. Let us extend this denition in the following way.
Denition 4.15. We say that f is eventually aperiodic whenever
the following property holds for every (n, k) N+ N: If A is a
measurable set such that for every x A there is 0 i k such that
f n+i (x) = f i (x), then (A) = 0.
It follows easily from the denition that an eventually periodic
map is aperiodic. The converse is true for invertible maps but not
in general (e.g. the constant map f (x) = c where c is a measurable
point of zero mass).
With this denition we can state the following result.
Theorem 4.16. Every measure-sensitive map is eventually aperiodic
(and so aperiodic).

70

[CAP. 4: MEASURE-SENSITIVE MAPS

Proof. Let f be a measure-sensitive map of X. Take (n, k) N+ N


and a measurable set A such that for every x A there is 0 i k
such that f n+i (x) = f i (x). Then,
A

k


f i (F ix(f n )),

(4.10)

i=0

where F ix(g) = {x X : g(x) = x} denotes the set of xed points


of a map g. Let P be a measure-sensitive partition of f . Then,
k+n m
(P ) is a countable partition. Fix x, y A. In particular
m=0 f

 k+n

m
f (P ) (x)
=
m=0

whence
y

 k+n



f

(P ) (x).

m=0

This together with (4.6) and (4.7) yields


f m (y) P (f m (x)),

0 m k + n.

(4.11)

But x, y A so (4.10) implies f i (x), f j (y) F ix(f n ) for some i, j


{0, , k}. We can assume that j i (otherwise we interchange the
roles of x and y in the argument below).
Now take m > k + n. Then, m > j + n so m j = pn + r for
some p N+ and some integer 0 r < n. Since 0 j + r < k + n
(for 0 j k and 0 r < n) one gets
f m (y) = f mj (f j (y))

=
=
=
(4.11)

f pn+r (f j (y))
f r (f pn (f j (y)))
f j+r (y)
P (f j+r (x)).

But
P (f j+r (x)) = P (f j+ri (f i (x)))

= P (f j+ri (f pn (f i (x))))
= P (f mi (f i (x)))
= P (f m (x))

71

[SEC. 4.4: APERIODICITY

so
f m (y) P (f m (x)),

m > k + n.

This together with (4.11) implies that f m (y) P (f m (x)) for all
m N whence y P (x). Consequently A P (x). As P is
measure-sensitive, Lemma 4.6 implies
(A ) = 0,

k+n


f i (P ).

i=0

On the other hand,

k+n
i=0

f i (P ) is a partition so


A=

and then (A) = 0 since


proof.

 k+n
i=0

k+n
i=0

(A )

f i (P )

f i (P ) is countable. This ends the

By Lemma 4.5 we have that, in non-atomic probability spaces, every measurable map carrying strong generators is measure-sensitive.
This motivates the question as to whether every measure-sensitive
map has a strong generator. We give a partial positive answer for
certain maps dened as follows. We say that f is countable to one
(mod 0) if f 1 (x) is countable for -a.e. x X. We say that f
is nonsingular if a measurable set A has measure zero if and only if
f 1 (A) also does. All measure-preserving maps are nonsingular. A
Lebesgue probability space is a complete measure space which is isomorphic to the completion of a standard probability space (c.f. [1],
[10]).
Corollary 4.17. The following properties are equivalent for nonsingular countable to one (mod 0) maps f on non-atomic Lebesgue
probability spaces:
1. f is measure-sensitive.
2. f is eventually aperiodic.
3. f is aperiodic.

72

[CAP. 4: MEASURE-SENSITIVE MAPS

4. f has a strong generator.


Proof. Notice that (1) (2) by Theorem 4.16 and (2) (3) follows
from the denitions. On the other hand, (3) (4) by a Parrys
Theorem (c.f. [69], [71], [70]) while (4) (1) by Lemma 4.5.
Denote by F ix(g) = {x X : g(x) = x} the set of xed points of
a mapping g.
Corollary 4.18. If f k = f for some integer k 2, then f is not
measure-sensitive.
Proof. Suppose by contradiction that it does. Then, f is eventually aperiodic by Theorem 4.16. On the other hand, if x X
then f k (x) = f (x) so f k1 (f k (x)) = f k1 (f (x)) = f k (x) therefore f k (x) F ix(f k1 ) whence X f k (F ix(f k1 )). But since f
is eventually aperiodic, n = k 1 N+ and X measurable we obtain
from the denition that (X) = 0 which is absurd. This ends the
proof.
Example 4.19. By Corollary 4.18 neither the identity f (x) = x nor
the constant map f (x) = c are measure-sensitive (for they satisfy
f 2 = f ). In particular, the converse of Theorem 4.16 is false for the
constant maps are eventually aperiodic but not measure-sensitive.
It is not dicult to prove that an ergodic measure-preserving map
of a non-atomic probability space is aperiodic. Then, Corollary 4.14
implies the well-known fact that all ergodic measure-preserving maps
with positive entropy on probability spaces are aperiodic. However,
using theorems 4.12 and 4.16 we obtain the following stronger result.
Corollary 4.20. All ergodic measure-preserving maps with positive
entropy on probability spaces are eventually aperiodic.
Now we study the following variant of aperiodicity introduced in
[41] p. 180.
Denition 4.21. We say that f is HS-aperiodic (2 ) whenever for
every measurable set of positive measure A and n N+ there is a
measurable subset B A such that (B \ f n (B)) > 0.
2 called

aperiodic in [41].

73

[SEC. 4.4: APERIODICITY

Notice that HS-aperiodicity implies the aperiodicity used in [48]


or [83] (for further comparisons see p. 88 in [56]).
On the other hand, a measurable map f is negative nonsingular
if (f 1 (A)) = 0 whenever A is a measurable set with (A) = 0.
Some consequences of the aperiodicity on negative nonsingular maps
in probability spaces are given in [56]. Observe that every measurepreserving map is negatively nonsingular.
Let us present two technical (but simple) results for later usage.
We call a measurable set A satisfying A f 1 (A) (mod 0) a positively invariant set (mod 0). For completeness we prove the following
property of these sets.
Lemma 4.22. If A is a positively invariant set (mod 0) of nite
measure of a negative nonsingular map f , then


(A)

= (A).

(4.12)

n=0

Proof. Since (A) = (A\f 1 (A))+(Af 1 (A)) and A is positively


invariant (mod 0) one has (A) = (A f 1 (A)), i.e.,


1


(A)

= (A).

n=0

Now suppose that m N+ satises




m


(A)

= (A).

n=0

Since

m+1



f

(A)

n=0

m


n=0


f

(A)



m


n=0


f

(A)


\f

m1

(A)

74

[CAP. 4: MEASURE-SENSITIVE MAPS

and


m


f n (A)

\ f m1 (A)

(f m (A) \ f m1 (A))

n=0

= (f m (A \ f 1 (A)))
= 0
because f is negative
nonsingular
and A is positively invariant (mod


m+1 n
(A) = (A). Therefore
0), one has
n=0 f


m



f n (A)

= (A),

m N,

(4.13)

n=0

by induction. On the other hand,

f n (A) =

n=0

and


m+1
n=0

f n (A)


f n (A)

n=0

m



f n (A)

m=0 n=0

m

f n (A). As (A) < we conclude that


m


(4.13)
n
= lim
f (A)
= lim (A) = (A)
n=0

n=0

proving (4.12).
We use the above lemma only in the proof of the proposition
below.
Proposition 4.23. Let P be a measure-sensitive partition of a negative nonsingular map f . Then, no P with positive nite measure
is positively invariant (mod 0).
Proof. Suppose by contradiction that there is P with 0 < () <
which is positively invariant (mod 0). Taking A = in Lemma
4.22 we obtain



n
f () = ().
(4.14)

n=0

75

[SEC. 4.4: APERIODICITY


As () > 0 we conclude that n=0 f n ()
= , and so, there is x
n
such that f (x) for all n N. As P we obtain P (f n (x)) =
and so f n (P (f n (x))) = f n () for all n N. Using (4.6) we get
Pm (x) =

m


f n ().

n=0

Then, (4.5) yields


P (x) =


m=0

Pm (x) =

m


m=0 n=0

f n () =

f n ()

n=0

and so (P (x)) = () by (4.14). Then, () = 0 by Lemma


4.6 since P is measure-sensitive which is absurd. This contradiction
proves the result.
We also need the following lemma resembling a well-known property of the expansive maps.
Lemma 4.24. If k N+ , then f is measure-sensitive if and only if
f k is.
f
Proof. The notation P
(x) will indicate the dependence of P (x)
on f .
First of all suppose that f k is an measure-sensitive with measurefk
(x)) = 0 for all x X by Lemma
sensitive partition P . Then, (P
f
fk
f
(x) so (P
(x)) = 0 for
4.6. But by denition one has P (x) P
all x X. Therefore, f is measure-sensitive with measure-sensitive
partition P . Conversely, suppose that f is measure-sensitive with exk
pansivity constant P . Consider Q = i=0 f i (P ) which is a countk
able partition satisfying Q(x) = i=0 f i (P (f i (x))) by (4.7). Now,
k
take y Qf (x). In particular, y Q(x) hence f i (y) P (f i (x)) for
every 0 i k. Take n > k so n = pk + r for some nonnegative
k
integers p and 0 r < k. As y Qf (x) one has f pk (y) Q(f pk (x))
and then f n (y) = f pk+i (y) = f i (f pk (y)) P (f i (f pk (x)) = P (f n (x))
proving f n (y) P (f n (x)) for all n N. Then, y P (x) yielding
k
k
Qf (x) P (x). Thus (Qf (x)) = 0 for all x X by the equivalence (ii) in Lemma 4.6 since P is measure-sensitive. It follows that
f k is measure-sensitive with measure-sensitive partition Q.

76

[CAP. 4: MEASURE-SENSITIVE MAPS

With these denitions and preliminary results we obtain the following.


Theorem 4.25. Every measure-sensitive negative nonsingular map
in a probability space is HS-aperiodic.
Proof. Suppose by contradiction that there is a measure-sensitive
map f which is negative nonsingular but not HS-aperiodic. Then,
there are a measurable set of positive measure A and n N+ such
that (B \ f n (B)) = 0 for every measurable subset B A. It
follows that every measurable subset B A is positively invariant
(mod 0) with respect to f n . By Lemma 4.24 we can assume n = 1.
Now, let P be a measure-sensitive partition of f . Clearly, since
(A) > 0 there is P such that (A ) > 0. Taking = A we
obtain that is positively invariant (mod 0) with positive measure.
In addition, consider the new partition Q = (P \ {}) {, \ A}
which is clearly measure-sensitive (for P is). Since this partition also
carries a positively invariant (mod 0) member of positive measure
(say ) we obtain a contradiction by Proposition 4.23. The proof
follows.

4.5

Exercices

Exercice 4.26.

Find non-standard measure-sensitive probability spaces.

Exercice 4.27.

Prove that the condition (4.2) for a sequence of partitions to be


measure-sensitive is also necessary in probability spaces.

Exercice 4.28. Is the converse of Proposition 4.3 true among probability spaces,
namely, is every non-atomic probability space measure-sensitive?
Exercice 4.29.

Prove the assertion in Example 4.7.

Exercice 4.30.

Prove that if Pn is a measure-sensitive sequence of partitions


of a probability space (X, B, ), then limn h(f, Pn ) exists for every measurepreserving map f : X X. Prove that this limit may depend on the measuresensitive sequence Pn .

[SEC. 4.5: EXERCICES

Exercice 4.31.

77

Prove that every measurable map of a separable metric space

which is pairwise sensitive with respect to a Borel probability measure is measuresensitive with respect to . Find a counterexample for the converse of this statement.

Exercice 4.32.

Prove that every expansive map of a separable non-atomic metric

space is measure-sensitive with respect to any non-atomic Borel probability measure.

Bibliography
[1] Aaronson, J., An introduction to innite ergodic theory, Mathematical Surveys and Monographs, 50. American Mathematical
Society, Providence, RI, 1997.
[2] Adler, R., L., Symbolic dynamics and Markov partitions, Bull.
Amer. Math. Soc. (N.S.) 35 (1998), no. 1, 156.
[3] Adler, R., L., Konheim, A., G., McAndrew, M., H., Topological
entropy, Trans. Amer. Math. Soc. 114 (1965), 309319.
[4] Akin, E., The general topology of dynamical systems. Graduate Studies in Mathematics, 1. American Mathematical Society,
Providence, RI, 1993.
[5] Aoki, N., Hiraide, K., Topological theory of dynamical systems, Recent advances. North-Holland Mathematical Library,
52. North-Holland Publishing Co., Amsterdam, 1994.
[6] Arbieto, A., Morales, C.A., Some properties of positive entropy
maps, Ergodic Theory Dynam. Systems (to appear).
[7] Arbieto, A., Morales, C.A., Expansivity of ergodic measures with
positive entropy, arXiv:1110.5598v1 [math.DS] 25 Oct 2011.
[8] Auslander, J., Berg, K., A condition for zero entropy, Israel J.
Math. 69 (1990), no. 1, 5964.
[9] Barreira, L., Pesin, Y., Schmeling, J., Dimension and product
structure of hyperbolic measures, Ann. of Math. (2) 149 (1999),
no. 3, 755783.
79

80

BIBLIOGRAPHY

[10] Bowen, R., Entropy-expansive maps, Trans. Amer. Math. Soc.


164 (1972), 323331.
[11] Bowen, R., Some systems with unique equilibrium states, Math.
Systems Theory 8 (1974/75), no. 3, 193202.
[12] Bowen, R., Ruelle, D., The ergodic theory of Axiom A ows,
Invent. Math. 29 (1975), no. 3, 181202.
[13] Blanchard, F., Huang, W., Snoha, L., Topological size of scrambled sets Colloq. Math. 110 (2008), no. 2, 293361.
[14] Blanchard, F., Host, B., Ruette, S., Asymptotic pairs in positiveentropy systems, Ergodic Theory Dynam. Systems 22 (2002), no.
3, 671686.
[15] Blanchard, F., Glasner, E., Kolyada, S., Maass, A., On Li-Yorke
pairs, J. Reine Angew. Math. 547 (2002), 5168.
[16] Block, L., S., Coppel, W., A., Dynamics in one dimension, Lecture Notes in Math., 1513, Springer-Verlag, Berlin, 1992.
[17] Bogachev, V., I., Measure theory. Vol. II. Springer-Verlag,
Berlin, 2007.
[18] Brin, M., Katok, A., On local entropy, Geometric dynamics (Rio
de Janeiro, 1981), 3038, Lecture Notes in Math., 1007, Springer,
Berlin, 1983.
[19] Bryant, B., F., Expansive Self-Homeomorphisms of a Compact
Metric Space, Amer. Math. Monthly 69 (1962), no. 5, 386391.
[20] Bryant, B. F., Walters, P., Asymptotic properties of expansive
homeomorphisms, Math. Systems Theory 3 (1969), 6066.
[21] Buzzi, J., Fisher, T., Entropic stability beyond partial hyperbolicity, Preprint arXiv:1103.2707v1 [math.DS] 14 Mar 2011.
[22] Carrasco-Olivera, D., Morales, C.A., Expansive measures for
ows, Preprint 2013 (to appear).
[23] Cao, Y.,, Zhao, Y., Measure-theoretic pressure for subadditive
potentials, Nonlinear Anal. 70 (2009), no. 6, 22372247.

BIBLIOGRAPHY

81

[24] Cerminara, M., Sambarino, M., Stable and unstable sets of C 0


perturbations of expansive homeomorphisms of surfaces, Nonlinearity 12 (1999), no. 2, 321332.
[25] Cerminara, M., Lewowicz, J., Some open problems concerning
expansive systems, Rend. Istit. Mat. Univ. Trieste 42 (2010),
129141.
[26] Cheng, W-C., Forward generator for preimage entropy, Pacic
J. Math. 223 (2006), no. 1, 516.
[27] Cadre, B., Jacob, P., On pairwise sensitivity, J. Math. Anal.
Appl. 309 (2005), no. 1, 375382.
[28] Coven, E., M., Keane, M., Every compact metric space that supports a positively expansive homeomorphism is nite, Dynamics
& stochastics, 304305, IMS Lecture Notes Monogr. Ser., 48,
Inst. Math. Statist., Beachwood, OH, 2006.
[29] Das, R., Expansive self-homeomorphisms on G-spaces, Period.
Math. Hungar. 31 (1995), no. 2, 123130.
[30] Dai, X., Geng, X., Zhou, Z., Some relations between Hausdordimensions and entropies, Sci. China Ser. A 41 (1998), no. 10,
10681075.
[31] Du, B-S. Every chaotic interval map has a scrambled set in the
recurrent set, Bull. Austral. Math. Soc. 39 (1989), no. 2, 259264.
[32] Dydak, J., Hoand, C., S., An alternative denition of coarse
structures, Topology Appl. 155 (2008), no. 9, 10131021.
[33] Eisenberg, M., Expansive transformation semigroups of endomorphisms, Fund. Math. 59 (1966), 313321.
[34] Fathi, A., Expansiveness, hyperbolicity and Hausdor dimension, Comm. Math. Phys. 126 (1989), no. 2, 249262.
[35] Fedorenko, V., V., Smital, J., Maps of the interval Ljapunov stable on the set of nonwandering points, Acta Math. Univ. Comenian. (N.S.) 60 (1991), no. 1, 1114.

82

BIBLIOGRAPHY

[36] Furstenberg, H., The structure of distal ows, Amer. J. Math.


85 1963 477515.
[37] Goodman, T., N., T., Relating topological entropy and measure
entropy, Bull. London Math. Soc. 3 (1971), 176180.
[38] Goodman, T., N., T., Maximal measures for expansive homeomorphisms, J. London Math. Soc. (2) 5 (1972), 439444.
[39] Gottschalk, W., H., Hedlund, G., A., Topological dynamics,
American Mathematical Society Colloquium Publications, Vol.
36. American Mathematical Society, Providence, R. I., 1955.
[40] Hasselblatt, B., Katok, A., Introduction to the modern theory
of dynamical systems. With a supplementary chapter by Katok
and Leonardo Mendoza, Encyclopedia of Mathematics and its
Applications, 54. Cambridge University Press, Cambridge, 1995.
[41] Helmberg, G., Simons, F., H., Aperiodic transformations, Z.
Wahrscheinlichkeitstheorie und Verw. Gebiete 13 (1969), 180
190.
[42] Hiraide, K., Expansive homeomorphisms of compact surfaces are
pseudo-Anosov, Proc. Japan Acad. Ser. A Math. Sci. 63 (1987),
no. 9, 337338.
[43] Huang, W., Stable sets and -stable sets in positive-entropy systems, Comm. Math. Phys. 279 (2008), no. 2, 535557.
[44] Huang, W., Lu, P., Ye, X., Measure-theoretical sensitivity and
equicontinuity, Israel J. Math. 183 (2011), 233283.
[45] Huang, W., Ye, X., Devaneys chaos or 2-scattering implies LiYorkes chaos, Topology Appl. 117 (2002), no. 3, 259272.
[46] Jakobsen, J., F., Utz, W., R., The non-existence of expansive
homeomorphisms on a closed 2-cell, Pacic J. Math. 10 (1960),
13191321.
[47] Jimenez Lopez, V., Large chaos in smooth functions of zero topological entropy, Bull. Austral. Math. Soc. 46 (1992), no. 2, 271
285.

BIBLIOGRAPHY

83

[48] Jones, L., K.,; Krengel, U., On transformations without nite


invariant measure, Advances in Math. 12 (1974), 275295.
[49] Kato, H., Continuum-wise expansive homeomorphisms, Canad.
J. Math. 45 (1993), no. 3, 576598.
[50] Kato, H., Chaotic continua of (continuum-wise) expansive homeomorphisms and chaos in the sense of Li and Yorke, Fund. Math.
145 (1994), no. 3, 261279.
[51] Kato, H., Expansive homeomorphisms on surfaces with holes,
Special volume in memory of Kiiti Morita. Topology Appl. 82
(1998), no. 1-3, 267277.
[52] Katok, A., Lyapunov exponents, entropy and periodic orbits for

dieomorphisms, Inst. Hautes Etudes


Sci. Publ. Math. No. 51
(1980), 137173.
[53] Katok, A., Bernoulli dieomorphisms on surfaces, Ann. of Math.
(2) 110 (1979), no. 3, 529547.
[54] Kifer, Y., Large deviations, averaging and periodic orbits of dynamical systems, Comm. Math. Phys. 162 (1994), no. 1, 3346.
[55] Knowles, J. D., On the existence of non-atomic measures, Mathematika 14 (1967), 6267.
[56] Kopf, C., Negative nonsingular transformations, Ann. Inst. H.
Poincare Sect. B (N.S.) 18 (1982), no. 1, 81102.
[57] Kuchta, M., Characterization of chaos for continuous maps of the
circle, Comment. Math. Univ. Carolin. 31 (1990), no. 2, 383390.
[58] Ledrappier, F., Young, L.-S., The metric entropy of dieomorphisms. II. Relations between entropy, exponents and dimension,
Ann. of Math. (2) 122 (1985), no. 3, 540574.
[59] Lewowicz, J., Expansive homeomorphisms of surfaces, Bol. Soc.
Brasil. Mat. (N.S.) 20 (1989), no. 1, 113133.
[60] Li, T.,Y., Yorke, J., A., Period three implies chaos, Amer. Math.
Monthly 82 (1975), no. 10, 985992.

84

BIBLIOGRAPHY

[61] Ma
ne, R., Ergodic theory and dierentiable dynamics. Translated
from the Portuguese by Silvio Levy. Ergebnisse der Mathematik
und ihrer Grenzgebiete (3) [Results in Mathematics and Related
Areas (3)], 8. Springer-Verlag, Berlin, 1987.
[62] Ma
ne, R., Expansive homeomorphisms and topological dimension, Trans. Amer. Math. Soc. 252 (1979), 313319.
[63] Ma
ne, R., Expansive dieomorphisms, Dynamical systems
Warwick 1974 (Proc. Sympos. Appl. Topology and Dynamical
Systems, Univ. Warwick, Coventry, 1973/1974; presented to E.
C. Zeeman on his ftieth birthday), pp. 162174. Lecture Notes
in Math., Vol. 468, Springer, Berlin, 1975.
[64] Morales, C.A., Partition sensitivity for measurable maps, Math.
Bohem. (to appear).
[65] Morales, C., A., A generalization of expansivity, Discrete Contin.
Dyn. Syst. 32 (2012), 293301.
[66] Morales, C.A., Sirvent, V., Expansivity for measures on uniform
spaces, Preprint 2012 (to appear).
[67] OBrien, T., Expansive homeomorphisms on compact manifolds,
Proc. Amer. Math. Soc. 24 (1970), 767771.
[68] Parry, W., Zero entropy of distal and related transformations,
1968 Topological Dynamics (Symposium, Colorado State Univ.,
Ft. Collins, Colo., 1967) pp. 383389 Benjamin, New York.
[69] Parry, W., Aperiodic transformations and generators, J. London
Math. Soc. 43 (1968), 191194.
[70] Parry, W., Principal partitions and generators, Bull. Amer.
Math. Soc. 73 (1967), 307309.
[71] Parry, W., Generators and strong generators in ergodic theory,
Bull. Amer. Math. Soc. 72 (1966), 294296.
[72] Parthasarathy, K. R., Probability measures on metric spaces,
Probability and Mathematical Statistics, No. 3 Academic Press,
Inc., New York-London 1967.

BIBLIOGRAPHY

85

[73] Parthasarathy, K., R., Ranga Rao, R., Varadhan, S., R., S.,
On the category of indecomposable distributions on topological
groups, Trans. Amer. Math. Soc. 102 (1962), 200217.
[74] Reddy, W., The existence of expansive homeomorphisms on
manifolds, Duke Math. J. 32 (1965), 627632.
[75] Reddy, W., Pointwise expansion homeomorphisms, J. London
Math. Soc. (2) 2 (1970), 232236.
[76] Reddy, W., Robertson, L., Sources, sinks and saddles for expansive homeomorphisms with canonical coordinates, Rocky Mountain J. Math. 17 (1987), no. 4, 673681.
[77] Richeson, D., Wiseman, J., Positively expansive homeomorphisms of compact spaces, Int. J. Math. Math. Sci. (2004), no
53-56, 29072910.
[78] Sakai, K., Hyperbolic metrics of expansive homeomorphisms,
Topology Appl. 63 (1995), no. 3, 263266.
[79] Schwartzman., S., On transformation groups, Ph.D. thesis, Yale
Univ., New Haven, CT, 1952.
[80] Sears, M., Expansive self-homeomorphisms of the Cantor set,
Math. Systems Theory 6 (1972), 129132.
[81] Sindelarova, P., A counterexample to a statement concerning
Lyapunov stability Acta Math. Univ. Comenianae 70 (2001),
265268.
[82] Smital, J., Chaotic functions with zero topological entropy,
Trans. Amer. Math. Soc. 297 (1986), no. 1, 269282.
[83] Steele, J., M., Covering nite sets by ergodic images, Canad.
Math. Bull. 21 (1978), no. 1, 8591.
[84] Takens, F.,Verbitski, E., Multifractal analysis of local entropies
for expansive homeomorphisms with specication, Comm. Math.
Phys. 203 (1999), no. 3, 593612.

86

BIBLIOGRAPHY

[85] Utz, W., R., Expansive mappings, Proceedings of the 1978 Topology Conference (Univ. Oklahoma, Norman, Okla., 1978), I.
Topology Proc. 3 (1978), no. 1, 221226 (1979).
[86] Utz, W., R., Unstable homeomorphisms, Proc. Amer. Math. Soc.
1 (1950), 769774.
[87] Vietez, J., L., Three-dimensional expansive homeomorphisms,
Dynamical systems (Santiago, 1990), 299323, Pitman Res.
Notes Math. Ser., 285, Longman Sci. Tech., Harlow, 1993.
[88] Vieitez, J., L., Expansive homeomorphisms and hyperbolic diffeomorphisms on 3-manifolds, Ergodic Theory Dynam. Systems
16 (1996), no. 3, 591622.
[89] Walters, P., An introduction to ergodic theory, Graduate Texts
in Mathematics, 79. Springer-Verlag, New York-Berlin, 1982.
[90] Williams, R., Some theorems on expansive homeomorphisms,
Amer. Math. Monthly 73 (1966), 854856.
[91] Williams, R., On expansive homeomorphisms, Amer. Math.
Monthly 76 (1969), 176178.
[92] Zhou, Z., L., Some equivalent conditions for self-mappings of a
circle, Chinese Ann. Math. Ser. A 12 (1991), suppl., 2227.

Index
countably-expansive, 2
Denjoy, 5
expansive, 1
with respect to (P), 1
h-expansive, 2
measure-expansive, 4
pointwise expansive, 38
proximal, 26

-generator, 14
positive, 44
Atom, 62
Class
stable, 46
Conjecture
Eckmann-Ruelle, 56
Constant
expansivity, 2
positive expansivity, 41
positively expansive, 46

Interval
wandering, 55
Manifold
closed, 6
Map
-isometry, 29
almost distal, 25
aperiodic, 69
eventually, 69
bijective
n-expansive, 32
n-expansive on A, 32
distal, 25
contably to one (mod 0),
71
continuous
Li-Yorke chaotic, 54
Denjoy, 34
entropy, 25

Dieomorphism
Axiom A, 6
Bernoulli, 57
Morse-Smale, 53
Entropy
Kolmogorov-Sinai, 67
zero, 2, 51
Exponent
Lyapunov, 56
Generator
strong, 64
Homeomorphism
-homeomorphism, 27
87

88
entropy of, 68
ergodic, 66
totally, 66
HS-aperiodic, 72
isometry, 4
Lyapunov stable on A, 47
measure-preserving, 66
measure-sensitive, 60
negative nonsingular, 73
nonsingular, 71
pairwise sensitive, 59
positively n-expansive, 32
positively n-expansive on
A, 32
positively expansive, 41
uniformly continuous, 5
uppersemicontinuous, 25
volume expanding, 45
Measure
almost expansive, 58
Borel, 2, 5, 6
dimension, 57
entropy, 49, 50
exact-dimensional, 56
expansive, 2
positively, 41
hyperbolic, 56, 59
Lebesgue, 4
maximal entropy, 39
pointwise expansive, 26
pullback, 5
space
measure-sensitive, 60
support, 11
Metric
n-discrete on A, 28
n-discrete on A
with constant , 28

INDEX

compact, 27
product, 13
restricted, 28
Number
Lebesgue, 15
Pair
asymptotic, 25, 46
Li-Yorke, 25
proximal, 25
Partition, 61
entropy, 67
measure-sensitive, 60, 64
sequence
increasing, 61
Lebesgue, 61
measure-sensitive, 61
Point
-isolated, 27
converging semiorbits, 15
heteroclinic, 20
periodic, 6
regular, 56
Principle
variational, 58
Set
-scrambled, 53
countable, 61
hyperbolic, 6
invariant, 5
negligible, 58
nonwandering, 6
positively invariant, 73
recurrent, 48
stable, 46
Space

INDEX

Lindelof, 4
measure, 61
non-atomic, 3, 62
probability, 2, 61
measure-sensitive, 61
metric
non-atomic, 3
Polish, 3
separable, 16
probability
Lebesgue, 61
probabiltity
Lebesgue, 71
Theorem
Bolzano-Weierstrass, 9
Brin-Katok, 49
Fubini, 43
Parry, 72
Poincare recurrence, 49
recurrence
Poincare, 11
Shannon-Breiman, 67
Topology
weak-*, 8, 9, 44

89